Work Characteristics and Employee Health

24
Work characteristics and employee health and well-being: Test of a model of healthy work organization Mark G. Wilson 1 *, David M. DeJoy 1 , Robert J. Vandenberg 2 , Hettie A. Richardson 3 and Allison L. McGrath 4 1 Workplace Health Group, Department of Health Promotion and Behavior, University of Georgia, USA 2 Workplace Health Group, Management Department, University of Georgia, USA 3 Management Department, Louisiana State University, USA 4 Raytheon Company, USA This paper presents an initial test and validation of a model of healthy work organi- zation. A questionnaire based on the proposed model was completed by 1,130 employees of a national retailer. The instrument measured 29 first-order constructs underlying the six higher-order domains of the model. The overall model fit and relationships among the second-order factors were examined using AMOS structural equation-modelling procedures. The structural analyses presented here support the proposed model. An acceptable overall fit was demonstrated, and all second-order, and second- to first-order, relationships were significant. Employees’ perceptions of their organization affect their perception of the climate, which impacts the way people relate to their job and see their future in the organization, ultimately impacting their work adjustment, health and well-being. This model has implications for both research and practice. Work organization has emerged as a priority topic among those who study the work- health relationship. Work organization generally refers to the way work processes are structured and managed, such as job design, scheduling, management, organizational characteristics, and policies and procedures (NIOSH, 1996). Inherent in this definition is the idea that the structure and fabric of the organization, and how it functions, can have a wide-ranging impact on the health and well-being of employees, and ultimately the effectiveness of the organization itself. The term ‘healthy work organization’ is a logical extension of work organization and presupposes that it should be possible to distinguish healthy from unhealthy work systems (e.g., Cooper & Williams, 1994; *Correspondence should be addressed to Dr Mark G. Wilson, Workplace Health Group, Department of Health Promotion and Behavior, University of Georgia, Athens, GA 30602, USA (e-mail: [email protected]). Journal of Occupational and Organizational Psychology (2004), 77, 565–588 © 2004 The British Psychological Society www.bps.org.uk 565

Transcript of Work Characteristics and Employee Health

Page 1: Work Characteristics and Employee Health

Work characteristics and employee health andwell-being Test of a model of healthy workorganization

Mark G Wilson1 David M DeJoy1 Robert J Vandenberg2Hettie A Richardson3 and Allison L McGrath4

1Workplace Health Group Department of Health Promotion and BehaviorUniversity of Georgia USA

2Workplace Health Group Management Department University of GeorgiaUSA

3Management Department Louisiana State University USA4Raytheon Company USA

This paper presents an initial test and validation of a model of healthy work organi-zation A questionnaire based on the proposed model was completed by 1130employees of a national retailer The instrument measured 29 first-order constructsunderlying the six higher-order domains of the model The overall model fit andrelationships among the second-order factors were examined using AMOS structuralequation-modelling procedures The structural analyses presented here support theproposed model An acceptable overall fit was demonstrated and all second-orderand second- to first-order relationships were significant Employeesrsquo perceptions oftheir organization affect their perception of the climate which impacts the waypeople relate to their job and see their future in the organization ultimately impactingtheir work adjustment health and well-being This model has implications for bothresearch and practice

Work organization has emerged as a priority topic among those who study the work-health relationship Work organization generally refers to the way work processes arestructured and managed such as job design scheduling management organizationalcharacteristics and policies and procedures (NIOSH 1996) Inherent in this definitionis the idea that the structure and fabric of the organization and how it functions canhave a wide-ranging impact on the health and well-being of employees and ultimatelythe effectiveness of the organization itself The term lsquohealthy work organizationrsquo is alogical extension of work organization and presupposes that it should be possibleto distinguish healthy from unhealthy work systems (eg Cooper amp Williams 1994

Correspondence should be addressed to Dr Mark G Wilson Workplace Health Group Department of Health Promotionand Behavior University of Georgia Athens GA 30602 USA (e-mail mwilsoncoeugaedu)

Journal of Occupational and Organizational Psychology (2004) 77 565ndash588copy 2004 The British Psychological Society

wwwbpsorguk

565

Cox 1988 Danna amp Griffin 1999 Landsbergis 2003 Lowe Schellenberg amp Shannon2003 NIOSH 2002 Smith Kaminstein amp Makadok 1995 Sparks Faragher amp Cooper2001)

In many respects healthy work organization represents the convergence of researchand thinking in several related areas of workplace research For example researchersin the human resourcesorganizational development area have focused on identifyingthe traits of healthy companies or organizations (eg Delery amp Shaw 2001 FerrisHochwarter Buckley Harrell-Cook amp Frink 1999 Jaffe 1995 Ostroff amp Bowen 2000Rosen amp Berger 1991) and exploring the characteristics of high-performance worksystems (eg Arthur 1994 Huselid 1995) Researchers in job stress have longbeen interested in delineating the job and organizational attributes that characterizehealthy or low-stress work environments But increasingly attention has shiftedtoward organizational or contextual factors in diagnosing and remedying the causes ofstress within organizations (eg Cartwright Cooper amp Murphy 1995 Cooper ampCartwright 1994 Lindstrom 1994 Peterson amp Wilson 2002 Sparks et al 2001) Asimilar shift is evident among occupational safety and health researchers especially theexpanding literatures on climateculture factors (eg Cheyne Cox Oliver amp Tomas1998 Cox amp Howarth 1990 DeJoy Searcy Murphy amp Gershon 2000 Griffin amp Neal2000 Lowe et al 2003 NIOSH 2002) and organization- or systems-level error (egHofmann Jacobs amp Landy 1995 Landsbergis 2003 Pate-Cornell 1990 Reason 19901995) Finally health-promotion researchers have turned to socialndashecological modelsand integrative or multilevel programming models (eg DeJoy amp Southern 1993DeJoy amp Wilson 2003 Pelletier 1984 Stokols 1992) and have shown an increasedinterest in examining covariations between employee and organizational outcomes(eg Goetzel Jacobson Aldana Vardell amp Yee 2000 Ozminkowski et al 1999)

Expanding the workndashhealth relationshipThe common thread in all of this work is the need to expand the workndashhealthrelationship beyond the immediate jobndashworker interaction and to provide a moresystematic accounting of macro-organizational influences (eg Cox Leather amp Cox1990 Danna amp Griffin 1999 DeJoy amp Southern 1993 Sauter Lim amp Murphy 1996Shannon Robson amp Sale 2001 Smith et al 1995) Although there has been consider-able discussion of healthy work organization there have been relatively few attemptsto develop or test actual models of healthy work organization Cox and colleagues(1990) and DeJoy and Southern (1993) are representative of some of the earlier workon this topic Cox and associates were among the first to argue that the study of workand health needs to be expanded to include the organizational context They identifiedthree primary sources of work demands the work itself the tools and technologiesused in the work and the socialndashorganizational and physical environments in whichthe work is performed DeJoy and Southern outlined a conceptual model consisting ofthree sets of nested factors beginning with the immediate workerndashjob interface andmoving outward to include the more inclusive socialndashorganizational environmentand finally to socialndashcultural factors and other extra-organizational influences Morerecently Danna and Griffin (1999) proposed an antecedent-consequences model thatfeatured three sets of antecedent factors work setting (primarily safety and healthrisks) personality traits and occupational stress factors In this model occupationalstress factors encompass both job demands as well as broader organizationalcharacteristics such as climate and career development opportunities

566 Mark G Wilson et al

Very few researchers have tried to test their conceptualizations of healthy workorganization In one preliminary study Sauter and colleagues (1996) examined thecontribution of three sets of organizational factors management practices organi-zational cultureclimate and organizational values Using previously collected data onapproximately 5000 employees of a large manufacturing concern separate regressionanalyses were performed for each category of factors Measures of perceived organi-zational effectiveness and perceived stress served as dependent variables From theregression coefficients presented the strongest results were found for continuousimprovement and career-development practices climate dimensions involving conflictresolution and fostering a sense of belonging and organizational values related toembracing technology and facilitating employee growth and development

Smith and colleagues (1995) analysed data from a large service organization andexamined five organizational factors organizationndashperson balance organizationaltreatment discrimination decision-making climate and quality of supervision Usingregression techniques and a health symptom index as the dependent measuredemographic factors accounted for about 4 of the variance and the addition oforganizationndashperson balance organizational treatment and discrimination brought thetotal explained variance to 20 Lowe and colleagues (Lowe et al 2003) sought toidentify the correlates of employeesrsquo perceptions of healthy work organization In thiscross-sectional study of 2500 Canadian workers psychosocial factors especially thosepertaining to interpersonal relations communications and social support were foundto be key ingredients in delineating a healthy workplace

In spite of the limited empirical data interest in healthy work organization continuesto grow For example basic aspects of healthy work organization and broad-basedassessment of organizational influences have shown up in recent reviews of researchon occupational safety and health issues (eg Shannon et al 2001) in studies ofemployee burnout (eg Demerouti Bakker Nachreiner amp Schaufeli 2001) and in thecurrent debate over health-care quality and patient safety (eg Sainfort Karsh Booskeamp Smith 2001) What is clearly lacking at this point is a direct and systematic test of acomprehensive model of healthy work organization A frequent lament in previousdiscussions of healthy work organization is that progress has been severely hamperedby the lack of comprehensive and testable models (eg Cox et al 1990 Danna ampGriffin 1999 DeJoy amp Wilson 2003 NIOSH 2002 Sauter et al 1996 Smith et al1995) Indeed relevant work has been largely piecemeal and typically confined to testsof individual job or work factors and narrowly defined employee outcomes Alsomuch of this research has been compartmentalized in various subspecializations suchas job stress or organizational design and this has served to impede cross-fertilizationeven between closely related areas of inquiry (DeJoy amp Wilson 2003 VandenbergPark DeJoy Wilson amp Griffin-Blake 2002) Although considerable empirical supportcan be found for virtually all of the linkages featured in the previous conceptualizationsof healthy work organization there have been virtually no attempts to systematicallytest comprehensive models of healthy work organization The model proposed belowseeks to capture the themes inherent in previous work and provide a framework forempirical testing The following working definition of healthy work organizationguided model development A healthy organization is one characterized by intentionalsystematic and collaborative efforts to maximize employee well-being and productiv-ity by providing well-designed and meaningful jobs a supportive socialndashorganizationalenvironment and accessible and equitable opportunities for career and workndashlifeenhancement

Healthy work organization 567

Proposed theoretical model of healthy work organizationThe model shown in Fig 1 attempts to incorporate the core aspects of healthy workorganization The model conceptualizes healthy work organization as consisting of sixinterrelated components Reviewing the diverse literature discussed earlier there issurprising agreement on the basic content domains of healthy work organizationIn one form or another three rather distinct domains of work life stand out andconsequently play an integral role in the model Job design emphasizes employeeslsquoindividual perceptions of their immediate work tasks Organizational climateemphasizes the social and interpersonal aspects of the work situation while job futureconcentrates on job security equity and career developments In one way or anotherprevious models of healthy work organization have made reference to these threegeneral domains

The literature on job stress and health provides much of the basis for identifyingspecific dimensions within the job-design component (Lindstrom 1994 Parker ampWahl 1998 Sauter Murphy amp Hurrell 1990 Warr 1994) Based on this literature sixdimensions were selected to represent the job design domain workload controlautonomy job content role clarity environmental conditions and work scheduleThere is less consensus for delineating the dimensions of organizational climate butorganizational or management support appears to play a key role in shaping theclimate of the organization (Ribisl amp Reischl 1993) Communication and opportunitiesfor meaningful participation are also frequently highlighted (James amp McIntyre 1996)We selected the following dimensions to comprise organizational climate organi-zational support coworker support participation and involvement communi-cation and health and safety climate Job Future represents a multidimensionaldomain that is more complex than simple job security (Sauter et al 1990) Careerdevelopment organizational fairness and work-life balance are important aspects oftodayrsquos employment relationship The dimensions chosen to represent this domain

Figure 1 Analytical model indicating the hypothesized associations among only the latent variablesAll variables except those with italicized labels were second-order latent variables Given that theywere operationalized with responses to single items the exceptions were treated as first-orderindicator variables in the analysis

568 Mark G Wilson et al

include job security pay and promotion equity learning opportunities and flexiblework arrangements

The remaining components include core organizational attributes psychologicalwork adjustment and employee health and well-being Almost without exceptionprevious writing on healthy work organization has emphasized the importance oforganization-level action in creating and maintaining healthy work organizationsThe core organizational attributes seek to delve into the culture and leadership orienta-tion of the organization at three different levels (Sauter et al 1996 Schien 1990)Presumably well-designed jobs supportive organizational climates and positive careerdevelopment options exist largely as a result of the visible policies established andactions taken by the leaders of the organization It also follows that policies andprocedures derive from the deeper beliefs and value structures of the organizationPsychological work adjustment underscores the importance of subjective evaluationand individual meaning in understanding the effects of various job and organizationalfactors on employee health and well-being (Lindstrom 1994) Four dimensions wereselected to represent employee work adjustment job satisfaction organizationalcommitment psychological empowerment and perceived job stress And finallypertinent to the present study Fig 1 includes several employee health and well-beingoutcomes

The healthy work organization concept centres on the premise that it should bepossible to identify the job and organizational characteristics of healthy organizationsand that such organizations should have healthier and more productive workers Asnoted earlier although several authors have tried to identify and describe the job andorganizational characteristics associated with healthy organizations (eg Cooper ampWilliams 1994 Danna amp Griffin 1999 Sauter et al 1996 Smith et al 1995) therehas been relatively little direct empirical research and no systematic analysis of amodel of healthy work organization conducted to date Moreover the specificdimensions comprising the major components of previous models have beeneither only minimally specified or not specified at all In the few instances when actualdata have been used to evaluate formulations of healthy work organization the datacame from existing data sets andor were collected for other or multiple purposesThe use of existing data sources typically necessitates partial tests andor the use ofsurrogate or approximate measures of important model constructs The purposeof this research was to conduct an initial test of a comprehensive healthy workorganization model

Materials and methods

Sample and data collectionThe sample consisted of 1130 employees from nine stores of one retail organizationThe stores were located in the south-eastern United States and varied in size from 150to 375 employees The 1130 respondents represented 53 of all employees in thenine locations Surveys were administered onsite during two consecutive workdays Questionnaire responses were completely anonymous and participation wasencouraged but voluntary As an incentive employees were given time on the clockto participate and provided a relatively quiet environment (a training room) inwhich to complete the survey Completed questionnaires were deposited inlocked storage boxes by the respondents to reinforce the confidentiality of theinformation

Healthy work organization 569

MeasuresOne instrument was used to operationalize all six higher-order components underlyingthe theoretical model For purposes of this study 194 items representing 29 first-orderconstructs were used from the instrument The development and validation of thequestionnaire consisted of a four-step process which included (1) selection ofcandidate measures from the scientific literature (2) content validation by threesubject matter experts from the National Institute for Occupational Safety and Health(3) exploratory factor analyses and (4) confirmatory factor analyses (details are avail-able from the authors by request) The description of the scales below follows theorder of the categories from left to right in Fig 1 The description begins with theexogenous variables (organizational attributes) and ends with the outcome variables(employee health and well-being) Unless otherwise indicated all measures werescaled on a 5-point Likert scale

Organizational attributesFrom the theoretical model core organizational attributes are conceptualized interms of three dimensions organizational values organizational beliefs and organi-zational policies and practices These three dimensions address the fundamentalimportance of organization-level action in creating and maintaining healthy work or-ganization

ValuesValues refer to the internalized normative beliefs which guide behaviour and desiredend-states within organizations (Rokeach 1979) While there are different lsquovalue setsrsquo(eg people production etc) the instrument focused on what OrsquoReilly Chatmanand Caldwell (1991) referred to as values with an employee orientation (eg toler-ance) The six-item measure asked respondents lsquothinking about your company ororganization as a whole how characteristic are each of the following traitsrsquo and theywere subsequently presented the six traits (alpha = 89)

BeliefsBeliefs entail employeesrsquo perceptions as to how the strongly the organization views itscommitment to and responsibility for employee health and well-being (Ribisl andReischl 1993 Sandroff Bradford and Gilligan 1990) An example from the nine-itemscale includes lsquoemployees should have a say in decisions that affect how they do theirjobsrsquo (α = 90)

Policies and practicesAn important attribute to reinforcing certain organizational values and beliefs is thebroad (ie organizational) policies and practices perceived by employees as the trueboundary conditions which guide their actions This study focused on three themesunderlying policies and practices which were based largely on the work of Jamiesonand OrsquoMara (1991) and of Vandenberg Richardson and Eastman (1999) All itemsregardless of theme asked respondents lsquoto what extent does your company or organi-zation as a whole have specific policies andor programs in place for rsquo The firsttheme was policies and practices for high involvement work practices and consisted of10 items (lsquoincorporating changes or innovations suggested by employees or employeegroupsrsquo α = 93) The second theme focused on policies and practices facilitating

570 Mark G Wilson et al

employeesrsquo abilities to balance work and nonwork issues (lsquooffering EAPs to helpemployees deal with stress family problems substance abuse etcrsquo α = 85) The thirdtheme was policies and practices reinforcing safety and health practices (lsquoprovidingapplicable occupational safety and health trainingrsquo α = 90) The scales for both thelast two themes consisted of four items each

Organizational climateThis component emphasized the perceptions of employees about their overall workenvironment particularly in terms of the climate for support communication andinvolvement Six dimensions were included in this component

Organizational supportOrganizational support was defined as the actions undertaken at the organizationallevel that encourage bolster or assist the employees in undertaking their tasks andresponsibilities Eisenberger Huntington Hutchison and Sowarsquos (1986) nine-itemglobal measure was used in this study (lsquothe organization really cares about mywell-beingrsquo α = 91)

Coworker supportThis variable focuses on the informal socialinterpersonal relationships that developamong peers Ribisl and Reischlrsquos scale (1993) was used to measure this construct (lsquomycoworkers care about me as a personrsquo α = 92)

Participation with others and with supervisorsParticipation in general refers to a climate in which employees are encouraged toinvolve themselves in some meaningful way with the people in the organization Thethree-item involvement with supervisors scale (Vroom 1959) included items such aslsquodo you feel you can influence decisions of your immediate supervisor regarding thingsabout which you are concernedrsquo (α = 77) The three-item involvement with othersscale (Caplan Cobb French Harrison amp Pinneau 1975) included items such as lsquoI takepart with others at my workplace in making decisions that affect mersquo (α = 88)

CommunicationThis is the extent to which employees see an effective information exchange withinthe organization The eight-item communication climate scale was adapted fromVandenberg and colleagues (1999) and included items such as lsquomanagement givesenough notice to employees before making changes in policies and proceduresrsquo(α = 86)

Safety and health climateThis involves the degree to which a climate for the safety and health of employees ispromoted in the work environment The seven-item scale used in the current studywas a version of the NIOSH Safety Climate Scale (DeJoy Murphy amp Gershon 1995)An item from the scale included lsquothere are no significant shortcuts taken whenworkplace safety and health are at stakersquo (α = 90)

Job designSix dimensions derived largely from reviews of the job stress literature (Cooper ampCartwright 1994 Lindstrom 1994 Sauter et al 1990) were included as part of the jobdesign component

Healthy work organization 571

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 2: Work Characteristics and Employee Health

Cox 1988 Danna amp Griffin 1999 Landsbergis 2003 Lowe Schellenberg amp Shannon2003 NIOSH 2002 Smith Kaminstein amp Makadok 1995 Sparks Faragher amp Cooper2001)

In many respects healthy work organization represents the convergence of researchand thinking in several related areas of workplace research For example researchersin the human resourcesorganizational development area have focused on identifyingthe traits of healthy companies or organizations (eg Delery amp Shaw 2001 FerrisHochwarter Buckley Harrell-Cook amp Frink 1999 Jaffe 1995 Ostroff amp Bowen 2000Rosen amp Berger 1991) and exploring the characteristics of high-performance worksystems (eg Arthur 1994 Huselid 1995) Researchers in job stress have longbeen interested in delineating the job and organizational attributes that characterizehealthy or low-stress work environments But increasingly attention has shiftedtoward organizational or contextual factors in diagnosing and remedying the causes ofstress within organizations (eg Cartwright Cooper amp Murphy 1995 Cooper ampCartwright 1994 Lindstrom 1994 Peterson amp Wilson 2002 Sparks et al 2001) Asimilar shift is evident among occupational safety and health researchers especially theexpanding literatures on climateculture factors (eg Cheyne Cox Oliver amp Tomas1998 Cox amp Howarth 1990 DeJoy Searcy Murphy amp Gershon 2000 Griffin amp Neal2000 Lowe et al 2003 NIOSH 2002) and organization- or systems-level error (egHofmann Jacobs amp Landy 1995 Landsbergis 2003 Pate-Cornell 1990 Reason 19901995) Finally health-promotion researchers have turned to socialndashecological modelsand integrative or multilevel programming models (eg DeJoy amp Southern 1993DeJoy amp Wilson 2003 Pelletier 1984 Stokols 1992) and have shown an increasedinterest in examining covariations between employee and organizational outcomes(eg Goetzel Jacobson Aldana Vardell amp Yee 2000 Ozminkowski et al 1999)

Expanding the workndashhealth relationshipThe common thread in all of this work is the need to expand the workndashhealthrelationship beyond the immediate jobndashworker interaction and to provide a moresystematic accounting of macro-organizational influences (eg Cox Leather amp Cox1990 Danna amp Griffin 1999 DeJoy amp Southern 1993 Sauter Lim amp Murphy 1996Shannon Robson amp Sale 2001 Smith et al 1995) Although there has been consider-able discussion of healthy work organization there have been relatively few attemptsto develop or test actual models of healthy work organization Cox and colleagues(1990) and DeJoy and Southern (1993) are representative of some of the earlier workon this topic Cox and associates were among the first to argue that the study of workand health needs to be expanded to include the organizational context They identifiedthree primary sources of work demands the work itself the tools and technologiesused in the work and the socialndashorganizational and physical environments in whichthe work is performed DeJoy and Southern outlined a conceptual model consisting ofthree sets of nested factors beginning with the immediate workerndashjob interface andmoving outward to include the more inclusive socialndashorganizational environmentand finally to socialndashcultural factors and other extra-organizational influences Morerecently Danna and Griffin (1999) proposed an antecedent-consequences model thatfeatured three sets of antecedent factors work setting (primarily safety and healthrisks) personality traits and occupational stress factors In this model occupationalstress factors encompass both job demands as well as broader organizationalcharacteristics such as climate and career development opportunities

566 Mark G Wilson et al

Very few researchers have tried to test their conceptualizations of healthy workorganization In one preliminary study Sauter and colleagues (1996) examined thecontribution of three sets of organizational factors management practices organi-zational cultureclimate and organizational values Using previously collected data onapproximately 5000 employees of a large manufacturing concern separate regressionanalyses were performed for each category of factors Measures of perceived organi-zational effectiveness and perceived stress served as dependent variables From theregression coefficients presented the strongest results were found for continuousimprovement and career-development practices climate dimensions involving conflictresolution and fostering a sense of belonging and organizational values related toembracing technology and facilitating employee growth and development

Smith and colleagues (1995) analysed data from a large service organization andexamined five organizational factors organizationndashperson balance organizationaltreatment discrimination decision-making climate and quality of supervision Usingregression techniques and a health symptom index as the dependent measuredemographic factors accounted for about 4 of the variance and the addition oforganizationndashperson balance organizational treatment and discrimination brought thetotal explained variance to 20 Lowe and colleagues (Lowe et al 2003) sought toidentify the correlates of employeesrsquo perceptions of healthy work organization In thiscross-sectional study of 2500 Canadian workers psychosocial factors especially thosepertaining to interpersonal relations communications and social support were foundto be key ingredients in delineating a healthy workplace

In spite of the limited empirical data interest in healthy work organization continuesto grow For example basic aspects of healthy work organization and broad-basedassessment of organizational influences have shown up in recent reviews of researchon occupational safety and health issues (eg Shannon et al 2001) in studies ofemployee burnout (eg Demerouti Bakker Nachreiner amp Schaufeli 2001) and in thecurrent debate over health-care quality and patient safety (eg Sainfort Karsh Booskeamp Smith 2001) What is clearly lacking at this point is a direct and systematic test of acomprehensive model of healthy work organization A frequent lament in previousdiscussions of healthy work organization is that progress has been severely hamperedby the lack of comprehensive and testable models (eg Cox et al 1990 Danna ampGriffin 1999 DeJoy amp Wilson 2003 NIOSH 2002 Sauter et al 1996 Smith et al1995) Indeed relevant work has been largely piecemeal and typically confined to testsof individual job or work factors and narrowly defined employee outcomes Alsomuch of this research has been compartmentalized in various subspecializations suchas job stress or organizational design and this has served to impede cross-fertilizationeven between closely related areas of inquiry (DeJoy amp Wilson 2003 VandenbergPark DeJoy Wilson amp Griffin-Blake 2002) Although considerable empirical supportcan be found for virtually all of the linkages featured in the previous conceptualizationsof healthy work organization there have been virtually no attempts to systematicallytest comprehensive models of healthy work organization The model proposed belowseeks to capture the themes inherent in previous work and provide a framework forempirical testing The following working definition of healthy work organizationguided model development A healthy organization is one characterized by intentionalsystematic and collaborative efforts to maximize employee well-being and productiv-ity by providing well-designed and meaningful jobs a supportive socialndashorganizationalenvironment and accessible and equitable opportunities for career and workndashlifeenhancement

Healthy work organization 567

Proposed theoretical model of healthy work organizationThe model shown in Fig 1 attempts to incorporate the core aspects of healthy workorganization The model conceptualizes healthy work organization as consisting of sixinterrelated components Reviewing the diverse literature discussed earlier there issurprising agreement on the basic content domains of healthy work organizationIn one form or another three rather distinct domains of work life stand out andconsequently play an integral role in the model Job design emphasizes employeeslsquoindividual perceptions of their immediate work tasks Organizational climateemphasizes the social and interpersonal aspects of the work situation while job futureconcentrates on job security equity and career developments In one way or anotherprevious models of healthy work organization have made reference to these threegeneral domains

The literature on job stress and health provides much of the basis for identifyingspecific dimensions within the job-design component (Lindstrom 1994 Parker ampWahl 1998 Sauter Murphy amp Hurrell 1990 Warr 1994) Based on this literature sixdimensions were selected to represent the job design domain workload controlautonomy job content role clarity environmental conditions and work scheduleThere is less consensus for delineating the dimensions of organizational climate butorganizational or management support appears to play a key role in shaping theclimate of the organization (Ribisl amp Reischl 1993) Communication and opportunitiesfor meaningful participation are also frequently highlighted (James amp McIntyre 1996)We selected the following dimensions to comprise organizational climate organi-zational support coworker support participation and involvement communi-cation and health and safety climate Job Future represents a multidimensionaldomain that is more complex than simple job security (Sauter et al 1990) Careerdevelopment organizational fairness and work-life balance are important aspects oftodayrsquos employment relationship The dimensions chosen to represent this domain

Figure 1 Analytical model indicating the hypothesized associations among only the latent variablesAll variables except those with italicized labels were second-order latent variables Given that theywere operationalized with responses to single items the exceptions were treated as first-orderindicator variables in the analysis

568 Mark G Wilson et al

include job security pay and promotion equity learning opportunities and flexiblework arrangements

The remaining components include core organizational attributes psychologicalwork adjustment and employee health and well-being Almost without exceptionprevious writing on healthy work organization has emphasized the importance oforganization-level action in creating and maintaining healthy work organizationsThe core organizational attributes seek to delve into the culture and leadership orienta-tion of the organization at three different levels (Sauter et al 1996 Schien 1990)Presumably well-designed jobs supportive organizational climates and positive careerdevelopment options exist largely as a result of the visible policies established andactions taken by the leaders of the organization It also follows that policies andprocedures derive from the deeper beliefs and value structures of the organizationPsychological work adjustment underscores the importance of subjective evaluationand individual meaning in understanding the effects of various job and organizationalfactors on employee health and well-being (Lindstrom 1994) Four dimensions wereselected to represent employee work adjustment job satisfaction organizationalcommitment psychological empowerment and perceived job stress And finallypertinent to the present study Fig 1 includes several employee health and well-beingoutcomes

The healthy work organization concept centres on the premise that it should bepossible to identify the job and organizational characteristics of healthy organizationsand that such organizations should have healthier and more productive workers Asnoted earlier although several authors have tried to identify and describe the job andorganizational characteristics associated with healthy organizations (eg Cooper ampWilliams 1994 Danna amp Griffin 1999 Sauter et al 1996 Smith et al 1995) therehas been relatively little direct empirical research and no systematic analysis of amodel of healthy work organization conducted to date Moreover the specificdimensions comprising the major components of previous models have beeneither only minimally specified or not specified at all In the few instances when actualdata have been used to evaluate formulations of healthy work organization the datacame from existing data sets andor were collected for other or multiple purposesThe use of existing data sources typically necessitates partial tests andor the use ofsurrogate or approximate measures of important model constructs The purposeof this research was to conduct an initial test of a comprehensive healthy workorganization model

Materials and methods

Sample and data collectionThe sample consisted of 1130 employees from nine stores of one retail organizationThe stores were located in the south-eastern United States and varied in size from 150to 375 employees The 1130 respondents represented 53 of all employees in thenine locations Surveys were administered onsite during two consecutive workdays Questionnaire responses were completely anonymous and participation wasencouraged but voluntary As an incentive employees were given time on the clockto participate and provided a relatively quiet environment (a training room) inwhich to complete the survey Completed questionnaires were deposited inlocked storage boxes by the respondents to reinforce the confidentiality of theinformation

Healthy work organization 569

MeasuresOne instrument was used to operationalize all six higher-order components underlyingthe theoretical model For purposes of this study 194 items representing 29 first-orderconstructs were used from the instrument The development and validation of thequestionnaire consisted of a four-step process which included (1) selection ofcandidate measures from the scientific literature (2) content validation by threesubject matter experts from the National Institute for Occupational Safety and Health(3) exploratory factor analyses and (4) confirmatory factor analyses (details are avail-able from the authors by request) The description of the scales below follows theorder of the categories from left to right in Fig 1 The description begins with theexogenous variables (organizational attributes) and ends with the outcome variables(employee health and well-being) Unless otherwise indicated all measures werescaled on a 5-point Likert scale

Organizational attributesFrom the theoretical model core organizational attributes are conceptualized interms of three dimensions organizational values organizational beliefs and organi-zational policies and practices These three dimensions address the fundamentalimportance of organization-level action in creating and maintaining healthy work or-ganization

ValuesValues refer to the internalized normative beliefs which guide behaviour and desiredend-states within organizations (Rokeach 1979) While there are different lsquovalue setsrsquo(eg people production etc) the instrument focused on what OrsquoReilly Chatmanand Caldwell (1991) referred to as values with an employee orientation (eg toler-ance) The six-item measure asked respondents lsquothinking about your company ororganization as a whole how characteristic are each of the following traitsrsquo and theywere subsequently presented the six traits (alpha = 89)

BeliefsBeliefs entail employeesrsquo perceptions as to how the strongly the organization views itscommitment to and responsibility for employee health and well-being (Ribisl andReischl 1993 Sandroff Bradford and Gilligan 1990) An example from the nine-itemscale includes lsquoemployees should have a say in decisions that affect how they do theirjobsrsquo (α = 90)

Policies and practicesAn important attribute to reinforcing certain organizational values and beliefs is thebroad (ie organizational) policies and practices perceived by employees as the trueboundary conditions which guide their actions This study focused on three themesunderlying policies and practices which were based largely on the work of Jamiesonand OrsquoMara (1991) and of Vandenberg Richardson and Eastman (1999) All itemsregardless of theme asked respondents lsquoto what extent does your company or organi-zation as a whole have specific policies andor programs in place for rsquo The firsttheme was policies and practices for high involvement work practices and consisted of10 items (lsquoincorporating changes or innovations suggested by employees or employeegroupsrsquo α = 93) The second theme focused on policies and practices facilitating

570 Mark G Wilson et al

employeesrsquo abilities to balance work and nonwork issues (lsquooffering EAPs to helpemployees deal with stress family problems substance abuse etcrsquo α = 85) The thirdtheme was policies and practices reinforcing safety and health practices (lsquoprovidingapplicable occupational safety and health trainingrsquo α = 90) The scales for both thelast two themes consisted of four items each

Organizational climateThis component emphasized the perceptions of employees about their overall workenvironment particularly in terms of the climate for support communication andinvolvement Six dimensions were included in this component

Organizational supportOrganizational support was defined as the actions undertaken at the organizationallevel that encourage bolster or assist the employees in undertaking their tasks andresponsibilities Eisenberger Huntington Hutchison and Sowarsquos (1986) nine-itemglobal measure was used in this study (lsquothe organization really cares about mywell-beingrsquo α = 91)

Coworker supportThis variable focuses on the informal socialinterpersonal relationships that developamong peers Ribisl and Reischlrsquos scale (1993) was used to measure this construct (lsquomycoworkers care about me as a personrsquo α = 92)

Participation with others and with supervisorsParticipation in general refers to a climate in which employees are encouraged toinvolve themselves in some meaningful way with the people in the organization Thethree-item involvement with supervisors scale (Vroom 1959) included items such aslsquodo you feel you can influence decisions of your immediate supervisor regarding thingsabout which you are concernedrsquo (α = 77) The three-item involvement with othersscale (Caplan Cobb French Harrison amp Pinneau 1975) included items such as lsquoI takepart with others at my workplace in making decisions that affect mersquo (α = 88)

CommunicationThis is the extent to which employees see an effective information exchange withinthe organization The eight-item communication climate scale was adapted fromVandenberg and colleagues (1999) and included items such as lsquomanagement givesenough notice to employees before making changes in policies and proceduresrsquo(α = 86)

Safety and health climateThis involves the degree to which a climate for the safety and health of employees ispromoted in the work environment The seven-item scale used in the current studywas a version of the NIOSH Safety Climate Scale (DeJoy Murphy amp Gershon 1995)An item from the scale included lsquothere are no significant shortcuts taken whenworkplace safety and health are at stakersquo (α = 90)

Job designSix dimensions derived largely from reviews of the job stress literature (Cooper ampCartwright 1994 Lindstrom 1994 Sauter et al 1990) were included as part of the jobdesign component

Healthy work organization 571

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 3: Work Characteristics and Employee Health

Very few researchers have tried to test their conceptualizations of healthy workorganization In one preliminary study Sauter and colleagues (1996) examined thecontribution of three sets of organizational factors management practices organi-zational cultureclimate and organizational values Using previously collected data onapproximately 5000 employees of a large manufacturing concern separate regressionanalyses were performed for each category of factors Measures of perceived organi-zational effectiveness and perceived stress served as dependent variables From theregression coefficients presented the strongest results were found for continuousimprovement and career-development practices climate dimensions involving conflictresolution and fostering a sense of belonging and organizational values related toembracing technology and facilitating employee growth and development

Smith and colleagues (1995) analysed data from a large service organization andexamined five organizational factors organizationndashperson balance organizationaltreatment discrimination decision-making climate and quality of supervision Usingregression techniques and a health symptom index as the dependent measuredemographic factors accounted for about 4 of the variance and the addition oforganizationndashperson balance organizational treatment and discrimination brought thetotal explained variance to 20 Lowe and colleagues (Lowe et al 2003) sought toidentify the correlates of employeesrsquo perceptions of healthy work organization In thiscross-sectional study of 2500 Canadian workers psychosocial factors especially thosepertaining to interpersonal relations communications and social support were foundto be key ingredients in delineating a healthy workplace

In spite of the limited empirical data interest in healthy work organization continuesto grow For example basic aspects of healthy work organization and broad-basedassessment of organizational influences have shown up in recent reviews of researchon occupational safety and health issues (eg Shannon et al 2001) in studies ofemployee burnout (eg Demerouti Bakker Nachreiner amp Schaufeli 2001) and in thecurrent debate over health-care quality and patient safety (eg Sainfort Karsh Booskeamp Smith 2001) What is clearly lacking at this point is a direct and systematic test of acomprehensive model of healthy work organization A frequent lament in previousdiscussions of healthy work organization is that progress has been severely hamperedby the lack of comprehensive and testable models (eg Cox et al 1990 Danna ampGriffin 1999 DeJoy amp Wilson 2003 NIOSH 2002 Sauter et al 1996 Smith et al1995) Indeed relevant work has been largely piecemeal and typically confined to testsof individual job or work factors and narrowly defined employee outcomes Alsomuch of this research has been compartmentalized in various subspecializations suchas job stress or organizational design and this has served to impede cross-fertilizationeven between closely related areas of inquiry (DeJoy amp Wilson 2003 VandenbergPark DeJoy Wilson amp Griffin-Blake 2002) Although considerable empirical supportcan be found for virtually all of the linkages featured in the previous conceptualizationsof healthy work organization there have been virtually no attempts to systematicallytest comprehensive models of healthy work organization The model proposed belowseeks to capture the themes inherent in previous work and provide a framework forempirical testing The following working definition of healthy work organizationguided model development A healthy organization is one characterized by intentionalsystematic and collaborative efforts to maximize employee well-being and productiv-ity by providing well-designed and meaningful jobs a supportive socialndashorganizationalenvironment and accessible and equitable opportunities for career and workndashlifeenhancement

Healthy work organization 567

Proposed theoretical model of healthy work organizationThe model shown in Fig 1 attempts to incorporate the core aspects of healthy workorganization The model conceptualizes healthy work organization as consisting of sixinterrelated components Reviewing the diverse literature discussed earlier there issurprising agreement on the basic content domains of healthy work organizationIn one form or another three rather distinct domains of work life stand out andconsequently play an integral role in the model Job design emphasizes employeeslsquoindividual perceptions of their immediate work tasks Organizational climateemphasizes the social and interpersonal aspects of the work situation while job futureconcentrates on job security equity and career developments In one way or anotherprevious models of healthy work organization have made reference to these threegeneral domains

The literature on job stress and health provides much of the basis for identifyingspecific dimensions within the job-design component (Lindstrom 1994 Parker ampWahl 1998 Sauter Murphy amp Hurrell 1990 Warr 1994) Based on this literature sixdimensions were selected to represent the job design domain workload controlautonomy job content role clarity environmental conditions and work scheduleThere is less consensus for delineating the dimensions of organizational climate butorganizational or management support appears to play a key role in shaping theclimate of the organization (Ribisl amp Reischl 1993) Communication and opportunitiesfor meaningful participation are also frequently highlighted (James amp McIntyre 1996)We selected the following dimensions to comprise organizational climate organi-zational support coworker support participation and involvement communi-cation and health and safety climate Job Future represents a multidimensionaldomain that is more complex than simple job security (Sauter et al 1990) Careerdevelopment organizational fairness and work-life balance are important aspects oftodayrsquos employment relationship The dimensions chosen to represent this domain

Figure 1 Analytical model indicating the hypothesized associations among only the latent variablesAll variables except those with italicized labels were second-order latent variables Given that theywere operationalized with responses to single items the exceptions were treated as first-orderindicator variables in the analysis

568 Mark G Wilson et al

include job security pay and promotion equity learning opportunities and flexiblework arrangements

The remaining components include core organizational attributes psychologicalwork adjustment and employee health and well-being Almost without exceptionprevious writing on healthy work organization has emphasized the importance oforganization-level action in creating and maintaining healthy work organizationsThe core organizational attributes seek to delve into the culture and leadership orienta-tion of the organization at three different levels (Sauter et al 1996 Schien 1990)Presumably well-designed jobs supportive organizational climates and positive careerdevelopment options exist largely as a result of the visible policies established andactions taken by the leaders of the organization It also follows that policies andprocedures derive from the deeper beliefs and value structures of the organizationPsychological work adjustment underscores the importance of subjective evaluationand individual meaning in understanding the effects of various job and organizationalfactors on employee health and well-being (Lindstrom 1994) Four dimensions wereselected to represent employee work adjustment job satisfaction organizationalcommitment psychological empowerment and perceived job stress And finallypertinent to the present study Fig 1 includes several employee health and well-beingoutcomes

The healthy work organization concept centres on the premise that it should bepossible to identify the job and organizational characteristics of healthy organizationsand that such organizations should have healthier and more productive workers Asnoted earlier although several authors have tried to identify and describe the job andorganizational characteristics associated with healthy organizations (eg Cooper ampWilliams 1994 Danna amp Griffin 1999 Sauter et al 1996 Smith et al 1995) therehas been relatively little direct empirical research and no systematic analysis of amodel of healthy work organization conducted to date Moreover the specificdimensions comprising the major components of previous models have beeneither only minimally specified or not specified at all In the few instances when actualdata have been used to evaluate formulations of healthy work organization the datacame from existing data sets andor were collected for other or multiple purposesThe use of existing data sources typically necessitates partial tests andor the use ofsurrogate or approximate measures of important model constructs The purposeof this research was to conduct an initial test of a comprehensive healthy workorganization model

Materials and methods

Sample and data collectionThe sample consisted of 1130 employees from nine stores of one retail organizationThe stores were located in the south-eastern United States and varied in size from 150to 375 employees The 1130 respondents represented 53 of all employees in thenine locations Surveys were administered onsite during two consecutive workdays Questionnaire responses were completely anonymous and participation wasencouraged but voluntary As an incentive employees were given time on the clockto participate and provided a relatively quiet environment (a training room) inwhich to complete the survey Completed questionnaires were deposited inlocked storage boxes by the respondents to reinforce the confidentiality of theinformation

Healthy work organization 569

MeasuresOne instrument was used to operationalize all six higher-order components underlyingthe theoretical model For purposes of this study 194 items representing 29 first-orderconstructs were used from the instrument The development and validation of thequestionnaire consisted of a four-step process which included (1) selection ofcandidate measures from the scientific literature (2) content validation by threesubject matter experts from the National Institute for Occupational Safety and Health(3) exploratory factor analyses and (4) confirmatory factor analyses (details are avail-able from the authors by request) The description of the scales below follows theorder of the categories from left to right in Fig 1 The description begins with theexogenous variables (organizational attributes) and ends with the outcome variables(employee health and well-being) Unless otherwise indicated all measures werescaled on a 5-point Likert scale

Organizational attributesFrom the theoretical model core organizational attributes are conceptualized interms of three dimensions organizational values organizational beliefs and organi-zational policies and practices These three dimensions address the fundamentalimportance of organization-level action in creating and maintaining healthy work or-ganization

ValuesValues refer to the internalized normative beliefs which guide behaviour and desiredend-states within organizations (Rokeach 1979) While there are different lsquovalue setsrsquo(eg people production etc) the instrument focused on what OrsquoReilly Chatmanand Caldwell (1991) referred to as values with an employee orientation (eg toler-ance) The six-item measure asked respondents lsquothinking about your company ororganization as a whole how characteristic are each of the following traitsrsquo and theywere subsequently presented the six traits (alpha = 89)

BeliefsBeliefs entail employeesrsquo perceptions as to how the strongly the organization views itscommitment to and responsibility for employee health and well-being (Ribisl andReischl 1993 Sandroff Bradford and Gilligan 1990) An example from the nine-itemscale includes lsquoemployees should have a say in decisions that affect how they do theirjobsrsquo (α = 90)

Policies and practicesAn important attribute to reinforcing certain organizational values and beliefs is thebroad (ie organizational) policies and practices perceived by employees as the trueboundary conditions which guide their actions This study focused on three themesunderlying policies and practices which were based largely on the work of Jamiesonand OrsquoMara (1991) and of Vandenberg Richardson and Eastman (1999) All itemsregardless of theme asked respondents lsquoto what extent does your company or organi-zation as a whole have specific policies andor programs in place for rsquo The firsttheme was policies and practices for high involvement work practices and consisted of10 items (lsquoincorporating changes or innovations suggested by employees or employeegroupsrsquo α = 93) The second theme focused on policies and practices facilitating

570 Mark G Wilson et al

employeesrsquo abilities to balance work and nonwork issues (lsquooffering EAPs to helpemployees deal with stress family problems substance abuse etcrsquo α = 85) The thirdtheme was policies and practices reinforcing safety and health practices (lsquoprovidingapplicable occupational safety and health trainingrsquo α = 90) The scales for both thelast two themes consisted of four items each

Organizational climateThis component emphasized the perceptions of employees about their overall workenvironment particularly in terms of the climate for support communication andinvolvement Six dimensions were included in this component

Organizational supportOrganizational support was defined as the actions undertaken at the organizationallevel that encourage bolster or assist the employees in undertaking their tasks andresponsibilities Eisenberger Huntington Hutchison and Sowarsquos (1986) nine-itemglobal measure was used in this study (lsquothe organization really cares about mywell-beingrsquo α = 91)

Coworker supportThis variable focuses on the informal socialinterpersonal relationships that developamong peers Ribisl and Reischlrsquos scale (1993) was used to measure this construct (lsquomycoworkers care about me as a personrsquo α = 92)

Participation with others and with supervisorsParticipation in general refers to a climate in which employees are encouraged toinvolve themselves in some meaningful way with the people in the organization Thethree-item involvement with supervisors scale (Vroom 1959) included items such aslsquodo you feel you can influence decisions of your immediate supervisor regarding thingsabout which you are concernedrsquo (α = 77) The three-item involvement with othersscale (Caplan Cobb French Harrison amp Pinneau 1975) included items such as lsquoI takepart with others at my workplace in making decisions that affect mersquo (α = 88)

CommunicationThis is the extent to which employees see an effective information exchange withinthe organization The eight-item communication climate scale was adapted fromVandenberg and colleagues (1999) and included items such as lsquomanagement givesenough notice to employees before making changes in policies and proceduresrsquo(α = 86)

Safety and health climateThis involves the degree to which a climate for the safety and health of employees ispromoted in the work environment The seven-item scale used in the current studywas a version of the NIOSH Safety Climate Scale (DeJoy Murphy amp Gershon 1995)An item from the scale included lsquothere are no significant shortcuts taken whenworkplace safety and health are at stakersquo (α = 90)

Job designSix dimensions derived largely from reviews of the job stress literature (Cooper ampCartwright 1994 Lindstrom 1994 Sauter et al 1990) were included as part of the jobdesign component

Healthy work organization 571

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 4: Work Characteristics and Employee Health

Proposed theoretical model of healthy work organizationThe model shown in Fig 1 attempts to incorporate the core aspects of healthy workorganization The model conceptualizes healthy work organization as consisting of sixinterrelated components Reviewing the diverse literature discussed earlier there issurprising agreement on the basic content domains of healthy work organizationIn one form or another three rather distinct domains of work life stand out andconsequently play an integral role in the model Job design emphasizes employeeslsquoindividual perceptions of their immediate work tasks Organizational climateemphasizes the social and interpersonal aspects of the work situation while job futureconcentrates on job security equity and career developments In one way or anotherprevious models of healthy work organization have made reference to these threegeneral domains

The literature on job stress and health provides much of the basis for identifyingspecific dimensions within the job-design component (Lindstrom 1994 Parker ampWahl 1998 Sauter Murphy amp Hurrell 1990 Warr 1994) Based on this literature sixdimensions were selected to represent the job design domain workload controlautonomy job content role clarity environmental conditions and work scheduleThere is less consensus for delineating the dimensions of organizational climate butorganizational or management support appears to play a key role in shaping theclimate of the organization (Ribisl amp Reischl 1993) Communication and opportunitiesfor meaningful participation are also frequently highlighted (James amp McIntyre 1996)We selected the following dimensions to comprise organizational climate organi-zational support coworker support participation and involvement communi-cation and health and safety climate Job Future represents a multidimensionaldomain that is more complex than simple job security (Sauter et al 1990) Careerdevelopment organizational fairness and work-life balance are important aspects oftodayrsquos employment relationship The dimensions chosen to represent this domain

Figure 1 Analytical model indicating the hypothesized associations among only the latent variablesAll variables except those with italicized labels were second-order latent variables Given that theywere operationalized with responses to single items the exceptions were treated as first-orderindicator variables in the analysis

568 Mark G Wilson et al

include job security pay and promotion equity learning opportunities and flexiblework arrangements

The remaining components include core organizational attributes psychologicalwork adjustment and employee health and well-being Almost without exceptionprevious writing on healthy work organization has emphasized the importance oforganization-level action in creating and maintaining healthy work organizationsThe core organizational attributes seek to delve into the culture and leadership orienta-tion of the organization at three different levels (Sauter et al 1996 Schien 1990)Presumably well-designed jobs supportive organizational climates and positive careerdevelopment options exist largely as a result of the visible policies established andactions taken by the leaders of the organization It also follows that policies andprocedures derive from the deeper beliefs and value structures of the organizationPsychological work adjustment underscores the importance of subjective evaluationand individual meaning in understanding the effects of various job and organizationalfactors on employee health and well-being (Lindstrom 1994) Four dimensions wereselected to represent employee work adjustment job satisfaction organizationalcommitment psychological empowerment and perceived job stress And finallypertinent to the present study Fig 1 includes several employee health and well-beingoutcomes

The healthy work organization concept centres on the premise that it should bepossible to identify the job and organizational characteristics of healthy organizationsand that such organizations should have healthier and more productive workers Asnoted earlier although several authors have tried to identify and describe the job andorganizational characteristics associated with healthy organizations (eg Cooper ampWilliams 1994 Danna amp Griffin 1999 Sauter et al 1996 Smith et al 1995) therehas been relatively little direct empirical research and no systematic analysis of amodel of healthy work organization conducted to date Moreover the specificdimensions comprising the major components of previous models have beeneither only minimally specified or not specified at all In the few instances when actualdata have been used to evaluate formulations of healthy work organization the datacame from existing data sets andor were collected for other or multiple purposesThe use of existing data sources typically necessitates partial tests andor the use ofsurrogate or approximate measures of important model constructs The purposeof this research was to conduct an initial test of a comprehensive healthy workorganization model

Materials and methods

Sample and data collectionThe sample consisted of 1130 employees from nine stores of one retail organizationThe stores were located in the south-eastern United States and varied in size from 150to 375 employees The 1130 respondents represented 53 of all employees in thenine locations Surveys were administered onsite during two consecutive workdays Questionnaire responses were completely anonymous and participation wasencouraged but voluntary As an incentive employees were given time on the clockto participate and provided a relatively quiet environment (a training room) inwhich to complete the survey Completed questionnaires were deposited inlocked storage boxes by the respondents to reinforce the confidentiality of theinformation

Healthy work organization 569

MeasuresOne instrument was used to operationalize all six higher-order components underlyingthe theoretical model For purposes of this study 194 items representing 29 first-orderconstructs were used from the instrument The development and validation of thequestionnaire consisted of a four-step process which included (1) selection ofcandidate measures from the scientific literature (2) content validation by threesubject matter experts from the National Institute for Occupational Safety and Health(3) exploratory factor analyses and (4) confirmatory factor analyses (details are avail-able from the authors by request) The description of the scales below follows theorder of the categories from left to right in Fig 1 The description begins with theexogenous variables (organizational attributes) and ends with the outcome variables(employee health and well-being) Unless otherwise indicated all measures werescaled on a 5-point Likert scale

Organizational attributesFrom the theoretical model core organizational attributes are conceptualized interms of three dimensions organizational values organizational beliefs and organi-zational policies and practices These three dimensions address the fundamentalimportance of organization-level action in creating and maintaining healthy work or-ganization

ValuesValues refer to the internalized normative beliefs which guide behaviour and desiredend-states within organizations (Rokeach 1979) While there are different lsquovalue setsrsquo(eg people production etc) the instrument focused on what OrsquoReilly Chatmanand Caldwell (1991) referred to as values with an employee orientation (eg toler-ance) The six-item measure asked respondents lsquothinking about your company ororganization as a whole how characteristic are each of the following traitsrsquo and theywere subsequently presented the six traits (alpha = 89)

BeliefsBeliefs entail employeesrsquo perceptions as to how the strongly the organization views itscommitment to and responsibility for employee health and well-being (Ribisl andReischl 1993 Sandroff Bradford and Gilligan 1990) An example from the nine-itemscale includes lsquoemployees should have a say in decisions that affect how they do theirjobsrsquo (α = 90)

Policies and practicesAn important attribute to reinforcing certain organizational values and beliefs is thebroad (ie organizational) policies and practices perceived by employees as the trueboundary conditions which guide their actions This study focused on three themesunderlying policies and practices which were based largely on the work of Jamiesonand OrsquoMara (1991) and of Vandenberg Richardson and Eastman (1999) All itemsregardless of theme asked respondents lsquoto what extent does your company or organi-zation as a whole have specific policies andor programs in place for rsquo The firsttheme was policies and practices for high involvement work practices and consisted of10 items (lsquoincorporating changes or innovations suggested by employees or employeegroupsrsquo α = 93) The second theme focused on policies and practices facilitating

570 Mark G Wilson et al

employeesrsquo abilities to balance work and nonwork issues (lsquooffering EAPs to helpemployees deal with stress family problems substance abuse etcrsquo α = 85) The thirdtheme was policies and practices reinforcing safety and health practices (lsquoprovidingapplicable occupational safety and health trainingrsquo α = 90) The scales for both thelast two themes consisted of four items each

Organizational climateThis component emphasized the perceptions of employees about their overall workenvironment particularly in terms of the climate for support communication andinvolvement Six dimensions were included in this component

Organizational supportOrganizational support was defined as the actions undertaken at the organizationallevel that encourage bolster or assist the employees in undertaking their tasks andresponsibilities Eisenberger Huntington Hutchison and Sowarsquos (1986) nine-itemglobal measure was used in this study (lsquothe organization really cares about mywell-beingrsquo α = 91)

Coworker supportThis variable focuses on the informal socialinterpersonal relationships that developamong peers Ribisl and Reischlrsquos scale (1993) was used to measure this construct (lsquomycoworkers care about me as a personrsquo α = 92)

Participation with others and with supervisorsParticipation in general refers to a climate in which employees are encouraged toinvolve themselves in some meaningful way with the people in the organization Thethree-item involvement with supervisors scale (Vroom 1959) included items such aslsquodo you feel you can influence decisions of your immediate supervisor regarding thingsabout which you are concernedrsquo (α = 77) The three-item involvement with othersscale (Caplan Cobb French Harrison amp Pinneau 1975) included items such as lsquoI takepart with others at my workplace in making decisions that affect mersquo (α = 88)

CommunicationThis is the extent to which employees see an effective information exchange withinthe organization The eight-item communication climate scale was adapted fromVandenberg and colleagues (1999) and included items such as lsquomanagement givesenough notice to employees before making changes in policies and proceduresrsquo(α = 86)

Safety and health climateThis involves the degree to which a climate for the safety and health of employees ispromoted in the work environment The seven-item scale used in the current studywas a version of the NIOSH Safety Climate Scale (DeJoy Murphy amp Gershon 1995)An item from the scale included lsquothere are no significant shortcuts taken whenworkplace safety and health are at stakersquo (α = 90)

Job designSix dimensions derived largely from reviews of the job stress literature (Cooper ampCartwright 1994 Lindstrom 1994 Sauter et al 1990) were included as part of the jobdesign component

Healthy work organization 571

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 5: Work Characteristics and Employee Health

include job security pay and promotion equity learning opportunities and flexiblework arrangements

The remaining components include core organizational attributes psychologicalwork adjustment and employee health and well-being Almost without exceptionprevious writing on healthy work organization has emphasized the importance oforganization-level action in creating and maintaining healthy work organizationsThe core organizational attributes seek to delve into the culture and leadership orienta-tion of the organization at three different levels (Sauter et al 1996 Schien 1990)Presumably well-designed jobs supportive organizational climates and positive careerdevelopment options exist largely as a result of the visible policies established andactions taken by the leaders of the organization It also follows that policies andprocedures derive from the deeper beliefs and value structures of the organizationPsychological work adjustment underscores the importance of subjective evaluationand individual meaning in understanding the effects of various job and organizationalfactors on employee health and well-being (Lindstrom 1994) Four dimensions wereselected to represent employee work adjustment job satisfaction organizationalcommitment psychological empowerment and perceived job stress And finallypertinent to the present study Fig 1 includes several employee health and well-beingoutcomes

The healthy work organization concept centres on the premise that it should bepossible to identify the job and organizational characteristics of healthy organizationsand that such organizations should have healthier and more productive workers Asnoted earlier although several authors have tried to identify and describe the job andorganizational characteristics associated with healthy organizations (eg Cooper ampWilliams 1994 Danna amp Griffin 1999 Sauter et al 1996 Smith et al 1995) therehas been relatively little direct empirical research and no systematic analysis of amodel of healthy work organization conducted to date Moreover the specificdimensions comprising the major components of previous models have beeneither only minimally specified or not specified at all In the few instances when actualdata have been used to evaluate formulations of healthy work organization the datacame from existing data sets andor were collected for other or multiple purposesThe use of existing data sources typically necessitates partial tests andor the use ofsurrogate or approximate measures of important model constructs The purposeof this research was to conduct an initial test of a comprehensive healthy workorganization model

Materials and methods

Sample and data collectionThe sample consisted of 1130 employees from nine stores of one retail organizationThe stores were located in the south-eastern United States and varied in size from 150to 375 employees The 1130 respondents represented 53 of all employees in thenine locations Surveys were administered onsite during two consecutive workdays Questionnaire responses were completely anonymous and participation wasencouraged but voluntary As an incentive employees were given time on the clockto participate and provided a relatively quiet environment (a training room) inwhich to complete the survey Completed questionnaires were deposited inlocked storage boxes by the respondents to reinforce the confidentiality of theinformation

Healthy work organization 569

MeasuresOne instrument was used to operationalize all six higher-order components underlyingthe theoretical model For purposes of this study 194 items representing 29 first-orderconstructs were used from the instrument The development and validation of thequestionnaire consisted of a four-step process which included (1) selection ofcandidate measures from the scientific literature (2) content validation by threesubject matter experts from the National Institute for Occupational Safety and Health(3) exploratory factor analyses and (4) confirmatory factor analyses (details are avail-able from the authors by request) The description of the scales below follows theorder of the categories from left to right in Fig 1 The description begins with theexogenous variables (organizational attributes) and ends with the outcome variables(employee health and well-being) Unless otherwise indicated all measures werescaled on a 5-point Likert scale

Organizational attributesFrom the theoretical model core organizational attributes are conceptualized interms of three dimensions organizational values organizational beliefs and organi-zational policies and practices These three dimensions address the fundamentalimportance of organization-level action in creating and maintaining healthy work or-ganization

ValuesValues refer to the internalized normative beliefs which guide behaviour and desiredend-states within organizations (Rokeach 1979) While there are different lsquovalue setsrsquo(eg people production etc) the instrument focused on what OrsquoReilly Chatmanand Caldwell (1991) referred to as values with an employee orientation (eg toler-ance) The six-item measure asked respondents lsquothinking about your company ororganization as a whole how characteristic are each of the following traitsrsquo and theywere subsequently presented the six traits (alpha = 89)

BeliefsBeliefs entail employeesrsquo perceptions as to how the strongly the organization views itscommitment to and responsibility for employee health and well-being (Ribisl andReischl 1993 Sandroff Bradford and Gilligan 1990) An example from the nine-itemscale includes lsquoemployees should have a say in decisions that affect how they do theirjobsrsquo (α = 90)

Policies and practicesAn important attribute to reinforcing certain organizational values and beliefs is thebroad (ie organizational) policies and practices perceived by employees as the trueboundary conditions which guide their actions This study focused on three themesunderlying policies and practices which were based largely on the work of Jamiesonand OrsquoMara (1991) and of Vandenberg Richardson and Eastman (1999) All itemsregardless of theme asked respondents lsquoto what extent does your company or organi-zation as a whole have specific policies andor programs in place for rsquo The firsttheme was policies and practices for high involvement work practices and consisted of10 items (lsquoincorporating changes or innovations suggested by employees or employeegroupsrsquo α = 93) The second theme focused on policies and practices facilitating

570 Mark G Wilson et al

employeesrsquo abilities to balance work and nonwork issues (lsquooffering EAPs to helpemployees deal with stress family problems substance abuse etcrsquo α = 85) The thirdtheme was policies and practices reinforcing safety and health practices (lsquoprovidingapplicable occupational safety and health trainingrsquo α = 90) The scales for both thelast two themes consisted of four items each

Organizational climateThis component emphasized the perceptions of employees about their overall workenvironment particularly in terms of the climate for support communication andinvolvement Six dimensions were included in this component

Organizational supportOrganizational support was defined as the actions undertaken at the organizationallevel that encourage bolster or assist the employees in undertaking their tasks andresponsibilities Eisenberger Huntington Hutchison and Sowarsquos (1986) nine-itemglobal measure was used in this study (lsquothe organization really cares about mywell-beingrsquo α = 91)

Coworker supportThis variable focuses on the informal socialinterpersonal relationships that developamong peers Ribisl and Reischlrsquos scale (1993) was used to measure this construct (lsquomycoworkers care about me as a personrsquo α = 92)

Participation with others and with supervisorsParticipation in general refers to a climate in which employees are encouraged toinvolve themselves in some meaningful way with the people in the organization Thethree-item involvement with supervisors scale (Vroom 1959) included items such aslsquodo you feel you can influence decisions of your immediate supervisor regarding thingsabout which you are concernedrsquo (α = 77) The three-item involvement with othersscale (Caplan Cobb French Harrison amp Pinneau 1975) included items such as lsquoI takepart with others at my workplace in making decisions that affect mersquo (α = 88)

CommunicationThis is the extent to which employees see an effective information exchange withinthe organization The eight-item communication climate scale was adapted fromVandenberg and colleagues (1999) and included items such as lsquomanagement givesenough notice to employees before making changes in policies and proceduresrsquo(α = 86)

Safety and health climateThis involves the degree to which a climate for the safety and health of employees ispromoted in the work environment The seven-item scale used in the current studywas a version of the NIOSH Safety Climate Scale (DeJoy Murphy amp Gershon 1995)An item from the scale included lsquothere are no significant shortcuts taken whenworkplace safety and health are at stakersquo (α = 90)

Job designSix dimensions derived largely from reviews of the job stress literature (Cooper ampCartwright 1994 Lindstrom 1994 Sauter et al 1990) were included as part of the jobdesign component

Healthy work organization 571

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 6: Work Characteristics and Employee Health

MeasuresOne instrument was used to operationalize all six higher-order components underlyingthe theoretical model For purposes of this study 194 items representing 29 first-orderconstructs were used from the instrument The development and validation of thequestionnaire consisted of a four-step process which included (1) selection ofcandidate measures from the scientific literature (2) content validation by threesubject matter experts from the National Institute for Occupational Safety and Health(3) exploratory factor analyses and (4) confirmatory factor analyses (details are avail-able from the authors by request) The description of the scales below follows theorder of the categories from left to right in Fig 1 The description begins with theexogenous variables (organizational attributes) and ends with the outcome variables(employee health and well-being) Unless otherwise indicated all measures werescaled on a 5-point Likert scale

Organizational attributesFrom the theoretical model core organizational attributes are conceptualized interms of three dimensions organizational values organizational beliefs and organi-zational policies and practices These three dimensions address the fundamentalimportance of organization-level action in creating and maintaining healthy work or-ganization

ValuesValues refer to the internalized normative beliefs which guide behaviour and desiredend-states within organizations (Rokeach 1979) While there are different lsquovalue setsrsquo(eg people production etc) the instrument focused on what OrsquoReilly Chatmanand Caldwell (1991) referred to as values with an employee orientation (eg toler-ance) The six-item measure asked respondents lsquothinking about your company ororganization as a whole how characteristic are each of the following traitsrsquo and theywere subsequently presented the six traits (alpha = 89)

BeliefsBeliefs entail employeesrsquo perceptions as to how the strongly the organization views itscommitment to and responsibility for employee health and well-being (Ribisl andReischl 1993 Sandroff Bradford and Gilligan 1990) An example from the nine-itemscale includes lsquoemployees should have a say in decisions that affect how they do theirjobsrsquo (α = 90)

Policies and practicesAn important attribute to reinforcing certain organizational values and beliefs is thebroad (ie organizational) policies and practices perceived by employees as the trueboundary conditions which guide their actions This study focused on three themesunderlying policies and practices which were based largely on the work of Jamiesonand OrsquoMara (1991) and of Vandenberg Richardson and Eastman (1999) All itemsregardless of theme asked respondents lsquoto what extent does your company or organi-zation as a whole have specific policies andor programs in place for rsquo The firsttheme was policies and practices for high involvement work practices and consisted of10 items (lsquoincorporating changes or innovations suggested by employees or employeegroupsrsquo α = 93) The second theme focused on policies and practices facilitating

570 Mark G Wilson et al

employeesrsquo abilities to balance work and nonwork issues (lsquooffering EAPs to helpemployees deal with stress family problems substance abuse etcrsquo α = 85) The thirdtheme was policies and practices reinforcing safety and health practices (lsquoprovidingapplicable occupational safety and health trainingrsquo α = 90) The scales for both thelast two themes consisted of four items each

Organizational climateThis component emphasized the perceptions of employees about their overall workenvironment particularly in terms of the climate for support communication andinvolvement Six dimensions were included in this component

Organizational supportOrganizational support was defined as the actions undertaken at the organizationallevel that encourage bolster or assist the employees in undertaking their tasks andresponsibilities Eisenberger Huntington Hutchison and Sowarsquos (1986) nine-itemglobal measure was used in this study (lsquothe organization really cares about mywell-beingrsquo α = 91)

Coworker supportThis variable focuses on the informal socialinterpersonal relationships that developamong peers Ribisl and Reischlrsquos scale (1993) was used to measure this construct (lsquomycoworkers care about me as a personrsquo α = 92)

Participation with others and with supervisorsParticipation in general refers to a climate in which employees are encouraged toinvolve themselves in some meaningful way with the people in the organization Thethree-item involvement with supervisors scale (Vroom 1959) included items such aslsquodo you feel you can influence decisions of your immediate supervisor regarding thingsabout which you are concernedrsquo (α = 77) The three-item involvement with othersscale (Caplan Cobb French Harrison amp Pinneau 1975) included items such as lsquoI takepart with others at my workplace in making decisions that affect mersquo (α = 88)

CommunicationThis is the extent to which employees see an effective information exchange withinthe organization The eight-item communication climate scale was adapted fromVandenberg and colleagues (1999) and included items such as lsquomanagement givesenough notice to employees before making changes in policies and proceduresrsquo(α = 86)

Safety and health climateThis involves the degree to which a climate for the safety and health of employees ispromoted in the work environment The seven-item scale used in the current studywas a version of the NIOSH Safety Climate Scale (DeJoy Murphy amp Gershon 1995)An item from the scale included lsquothere are no significant shortcuts taken whenworkplace safety and health are at stakersquo (α = 90)

Job designSix dimensions derived largely from reviews of the job stress literature (Cooper ampCartwright 1994 Lindstrom 1994 Sauter et al 1990) were included as part of the jobdesign component

Healthy work organization 571

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 7: Work Characteristics and Employee Health

employeesrsquo abilities to balance work and nonwork issues (lsquooffering EAPs to helpemployees deal with stress family problems substance abuse etcrsquo α = 85) The thirdtheme was policies and practices reinforcing safety and health practices (lsquoprovidingapplicable occupational safety and health trainingrsquo α = 90) The scales for both thelast two themes consisted of four items each

Organizational climateThis component emphasized the perceptions of employees about their overall workenvironment particularly in terms of the climate for support communication andinvolvement Six dimensions were included in this component

Organizational supportOrganizational support was defined as the actions undertaken at the organizationallevel that encourage bolster or assist the employees in undertaking their tasks andresponsibilities Eisenberger Huntington Hutchison and Sowarsquos (1986) nine-itemglobal measure was used in this study (lsquothe organization really cares about mywell-beingrsquo α = 91)

Coworker supportThis variable focuses on the informal socialinterpersonal relationships that developamong peers Ribisl and Reischlrsquos scale (1993) was used to measure this construct (lsquomycoworkers care about me as a personrsquo α = 92)

Participation with others and with supervisorsParticipation in general refers to a climate in which employees are encouraged toinvolve themselves in some meaningful way with the people in the organization Thethree-item involvement with supervisors scale (Vroom 1959) included items such aslsquodo you feel you can influence decisions of your immediate supervisor regarding thingsabout which you are concernedrsquo (α = 77) The three-item involvement with othersscale (Caplan Cobb French Harrison amp Pinneau 1975) included items such as lsquoI takepart with others at my workplace in making decisions that affect mersquo (α = 88)

CommunicationThis is the extent to which employees see an effective information exchange withinthe organization The eight-item communication climate scale was adapted fromVandenberg and colleagues (1999) and included items such as lsquomanagement givesenough notice to employees before making changes in policies and proceduresrsquo(α = 86)

Safety and health climateThis involves the degree to which a climate for the safety and health of employees ispromoted in the work environment The seven-item scale used in the current studywas a version of the NIOSH Safety Climate Scale (DeJoy Murphy amp Gershon 1995)An item from the scale included lsquothere are no significant shortcuts taken whenworkplace safety and health are at stakersquo (α = 90)

Job designSix dimensions derived largely from reviews of the job stress literature (Cooper ampCartwright 1994 Lindstrom 1994 Sauter et al 1990) were included as part of the jobdesign component

Healthy work organization 571

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 8: Work Characteristics and Employee Health

WorkloadAn employeersquos workload consists of the daily demands of the work situation Thisconstruct was measured with four items taken from a task demand scale developed byKlitzman House Israel and Mero (1990) An example of items included lsquoI am asked todo an excessive amount of workrsquo (α = 78)

ControlautonomyAutonomy is the degree to which the job provides substantial freedom independenceand discretion to the individual in scheduling the work and in determining theprocedures to be used in carrying it out The three-item scale contained in the JobDiagnostic Survey (Hackman amp Oldham 1975) was adapted for the current study (eglsquomy job permits me to decide on my own how to go about doing the workrsquo α = 77)

Job contentJob content is the extent to which the job is viewed as being meaningful valuable andworthwhile A six-item scale (House McMichael Wells Kaplan amp Landerman 1979)was used to measure this construct (lsquoI have an opportunity to develop my own specialskills and abilitiesrsquo α = 80)

Role clarityRole clarity is the extent to which an employeersquos work goals and responsibilities areclearly communicated and whether the individual understands the processes requiredto achieve these goals (Sawyer 1992) The four-item scale used to measure this con-struct was adapted from Rizzo House and Lirtzman (1970) A sample item includedlsquothere are clear planned goals and objectives for my jobrsquo (α = 82)

Environmental and physical work conditionsBased both on the work of Johansson Johnson and Hall (1991) and lsquowalk throughsrsquo ofthe storersquos work environment conducted by the authors two scales were created toassess the employeesrsquo working conditions The seven-item environmental conditionsscale encompassed employeesrsquo perceptions of the potential hazards found in theirimmediate work areas such as noise and poor lighting (α = 84) The five-item physicalwork demands scale assessed factors such as lifting and repetitive motion (α = 82)

Work schedulingThe scheduling of work time encompasses a variety of options such as full or part timefixed or rotating schedules and day versus other shift arrangements The five-itemwork schedule scale used in this study consisted of items from Morrow McElroy andElliott (1994) with additional items developed by the investigators to reflect schedulingissues unique to the participating company (lsquomy work hours are unpredictable fromone week to the nextrsquo α = 84)

Job futureIndustry restructuring globalization changes in employee benefits and other factorshave altered the traditional employerndashemployee relationship (Rousseau 1997) Toreflect these trends five dimensions were included in the job-future component (thediscussion of two procedural and distributive equity is collapsed for brevity)

572 Mark G Wilson et al

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 9: Work Characteristics and Employee Health

Job securityThis variable consists of the employeesrsquo perceptions about the likely continuity of theiremployment with the organization A five-item scale (Kuhnert Sims amp Lahey 1989)was used to measure this variable (lsquoI am afraid of losing my jobrsquo α = 79)

Procedural and distributive equityEquity refers in general to perceived fairness From its long and rich research historytwo principal forms of equity have emerged as important in the study of organizationsDistributive equity represents the perceived fairness attached to the amount ofrewards (eg merit pay increases) and their allocation along some performancecriterion The current study used the four-item distributive equity scale fromBavendam Boyer and Sorensen (1986) An example item is lsquoI am fairly rewardedconsidering my responsibilitiesrsquo (α = 95) Procedural equity entails the perceivedfairness of how the lsquorulesrsquo are applied across people The six-item procedural equityscale by Greenberg (1986) was used and included items such as lsquowhen pay andpromotion decisions are made all sides affected by the decisions have a saylsquo (α = 95)

Learning opportunitiesLearning opportunities entail employeesrsquo beliefs about available opportunities to learnnew skills or keep current skills updated Five items adapted from Vandenberg et al(1999) were used to assess this dimension (lsquoI am given a real opportunity to improvemy knowledge and skillsrsquo α = 90)

Flexible work arrangementsThis variable involves the extent to which job requirements limit employeesrsquo ability tofulfil various non-work obligations and activities Bohen and Viveros-Longrsquos (1981)six-item scale was used to measure this construct (lsquohow easy or difficult is it to arrangetime to do each of the following on a typical workdayrsquo with items including lsquogo to ahealth care appointmentrsquo and lsquorespond to the needs of your children or other familymembersrsquo α = 87)

Psychological work adjustmentThe five dimensions for this component were selected on the premise that satisfiedcommitted high efficacy and lsquolow stressrsquo employees contribute positively toorganizational effectiveness and are reflective of healthy work organization

Job satisfactionThis construct examines how satisfied employees are with their specific work situ-ation tasks demands and responsibilities The five-item scale (Hackman amp Oldham1975 1980) included items such as lsquogenerally speaking I am very satisfied with myjobrsquo (α = 81)

Organizational commitmentOrganizational commitment is the strength of employeesrsquo attachment to the companyThe nine-item version of the Organizational Commitment Questionnaire (MowdaySteers amp Porter 1979) was used to measure this construct (lsquoI am willing to put in agreat deal of effort beyond that normally expected in order to help this organization besuccessfulrsquo α = 92)

EfficacyEfficacy involves a personrsquos sense of mastery and confidence in their work roleAdapting Spreitzerrsquos (1995) perspective and operationalizations we utilized two forms

Healthy work organization 573

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 10: Work Characteristics and Employee Health

of efficacy Self-efficacy refers to a personrsquos own sense of confidence in their ability toeffectively work at their job This was measured with Spreitzerrsquos three-item scale (lsquoI amconfident about my ability to do my jobrsquo α = 81) The second form impact refers to apersonrsquos perception about their ability to meaningfully influence their workgroup orteam Again Spreitzerrsquos (1995) three-item measure was used (lsquomy impact on whathappens in my workgroup is largersquo α = 88)

Job stressJob stress focused on the employeersquos perceptions and reactions to stressors at workThe six-item scale was adapted from Cohen Kamarck and Mermelstein (1983) andincluded items such as lsquoin the last month how often have you been upset because ofsomething that happened unexpectedly at workrsquo (α = 88)

Employee health and well-beingFor this study this component included measures of employeesrsquo perceived generalhealth psychological health attendance behaviour (eg likelihood of turnover) andengagement in health risk behaviours (eg tobacco use)

Employee healthSelf-reported health was assessed using a single-item adapted from the SF-36 HealthSurvey (Ware amp Sherbourne 1992) The question asked respondents to rate theiroverall health from lsquopoorrsquo to lsquoexcellentrsquo

Psychological healthThree measures were used here The first two adapted from Ilfeld (1978) assesseddepressive symptoms and angerhostility Depressive symptoms were representedusing seven items (eg feel downhearted and blue feel lonely α = 86) and angerwas operationalized with four items (eg feel easily annoyed feel critical of othersα = 87) Respondents were asked to report how often they had experienced thesesymptoms or feelings during the past month using a four-point scale lsquoneverrsquo lsquoonce ina whilersquo lsquofairly oftenrsquo lsquovery oftenrsquo The third measure somatic symptoms includedseven symptom states generally associated with stress andor anxiety (eg headacheheart beating hard sweaty hands) Respondents indicated how often they hadexperienced these symptoms during the past month (α = 89) The term lsquopsycho-logical healthrsquo was used to denote a positive connotation as opposed to a negativeconnotation such as psychological dysfunction Low scores on these measures indicatehigh levels of psychological health

Health risk behavioursTwo measures were adapted from the US Centers for Disease Control and PreventionrsquosBehavioral Risk Factor Surveillance Survey to assess alcohol and tobacco use Thealcohol measure provided an overall measure of alcohol consumption during a 30-dayperiod It was the combination of responses to drinking frequency (number of daysalcohol consumed per month) multiplied by quantity (number of drinks typicallyconsumed on a day that alcoholic beverages were consumed) Tobacco use wasassessed using lsquoyesrsquo or lsquonorsquo responses to a question asking respondents whether theycurrently used tobacco (cigarettes cigarspipes or chewing or smokeless tobacco)

Attendance behavioursThis component was operationalized through employeesrsquo self-rated turnoverintentions and their frequency of absences Turnover intention was captured with a

574 Mark G Wilson et al

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 11: Work Characteristics and Employee Health

single item that queried respondents about their likelihood of leaving the companywithin the next 12 months The five response categories were in percent likelihoodintervals 0ndash20 (not likely) 21ndash40 (somewhat likely) and so forth Absenteeism wasalso assessed with a single item that asked respondents to indicate the number of daysthey had missed work during the previous 3 months (excluding vacations holidaysand other scheduled leave)

Analytical procedureThe analyses were guided by the two-stage process outlined in Gerbing and Anderson(1987) (1) tests of the measurement models followed by (2) the tests of the hypoth-esized associations among the constructs This process is based on the premise thatyou must know what you are measuring first before you can test any substantivehypothesis among the constructs represented through the measures This was particu-larly appropriate in our case because of the complexity resulting from the number ofdifferent scales used in the study All analyses were undertaken using the AMOSstructural equation-modelling program (Arbuckle amp Wothke 1999) and utilizing thevariancecovariance matrix Taking into consideration the complexity of the model sixfit indices were used chi-square goodness of fit test chi-square to degrees-of-freedomratio standardized root mean square residual (SRMSR) root mean square error ofapproximation (RMSEA Steiger 1990) TuckerndashLewis Index (TLI Tucker amp Lewis1973) and the Relative Noncentrality Index (RNI McDonald amp Marsh 1990) Theexpected maximization (full information maximum likelihood) multiple imputationprocedure (LISREL Version 85 Du Toit amp Du Toit 2001) was applied to impute themissing values As a point of reference only 4 of the total number of responses weremissing

Inferring support for the hypotheses was a two-step process The first step was toexamine the overall fit of the model to the data This is an omnibus test that in practicalterms asks whether or not the specification of the paths as conceptually supported is areasonable reflection of the theoretical process underlying the variables Assuming thatit is the second step is to examine the statistical significance of each of the hypoth-esized paths to infer direct support for each expectation It does not make sense toinfer the statistical significance of the hypothesized paths without first asking whetherthe model itself is reasonable

The same indices as above were used to infer fit Note however that the complexityis even greater at this point That is there is now a layer of second-order latentvariables on top of the layer of 34 first-order dimensions defined by the 194 itemsThis added complexity will have an impact on the fit indices that penalize modelcomplexity such as the TLI and RNI Therefore while the other fit indices may indicatea good fit the latter ones may be less than ideal

Results

The means standard deviations and correlations among the primary variables arereported in Table 1 The diagonal elements represent the internal consistency reliabilitycoefficients All were above 75

The seven tests of the measurement models are represented by the first seven rowsof Table 2 Looking first at the tests of the six major model components generally agood fit may be inferred among the different model components when the dimensionsof the components are tested relative to one another An exceptionally strong fit was

Healthy work organization 575

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 12: Work Characteristics and Employee Health

576 Mark G Wilson et al

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 13: Work Characteristics and Employee Health

Healthy work organization 577

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 14: Work Characteristics and Employee Health

578 Mark G Wilson et al

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 15: Work Characteristics and Employee Health

observed for the job future and health and well-being components In contrast the fitwas least strong but still acceptable for the work adjustment and organizationalattributes components of the model as indicated particularly by their chi-square todegrees-of-freedom ratios

The seventh row of Table 2 contains the fit indices for the test of the measurementmodel including all dimensions Three of the fit indices suggested a very strong fit thechi-square to degrees-of-freedom ratio the RMSEA and SRMSR In contrast the TLI andRNI fell somewhat below the 90 standard to infer strong fit The values for the TLI andRNI were not unexpected as these particular indices heavily penalize complex mod-els However the research literature provides no agreed upon standards for how muchof a drop in those two indices is required before observing true model misspecification(the aspect these indices are purportedly sensitive to) versus a sensitivity of theseindices to anomalous study characteristics (model complexity in this case) Given thatthe drop was not dramatic (to 80 or less) and that the other fit indices indicated animproved fit relative to the other models it was concluded that the measurementmodel with all 194 items and 34 dimensions had a strong fit In practical terms thismeant that each dimension was adequately represented through its respective item setand that there was not enough overlap between dimensions to be concerned aboutmisspecifications due to multicollinearity or conceptual redundancy

The last row in Table 2 provides the fit indices for the tests of hypothesizedassociations among the model components presented in Figure 1 As indicated therethe addition of 523 more degrees of freedom (representing the difference in df fromthe lsquobigrsquo measurement model to the hypothesized one) had little impact on the valuesof the fit indices relative to those observed from testing the lsquobigrsquo measurement model

Healthy work organization 579

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 16: Work Characteristics and Employee Health

The increased complexity had as expected the biggest impact on the TLI and RNI Allin all while taking into consideration the complexity of the hypothesized model itgenerally fitted the data and thus there was little concern that the model was mis-specifying the relationships among the variables As such the path coefficients couldbe meaningfully interpreted

In this vein and turning now to the top two-thirds of Table 3 the results indicatethat the specification of the second-order variables was very successful In brief all ofthe first-order dimensions had statistically significant loadings on their respectivesecond-order factors and the signs of the obtained estimates were all in the expecteddirections Strong support therefore was observed for specifying a common over-riding second-order latent construct for each set of the first-order dimensions That isthere was a systematic source of common variation across each of the dimensionswithin a set after accounting for the systematic variance within a dimension and thatwould have normally been pushed into the error term if the tests of hypotheses hadonly been conducted at the first-order level (ie the traditional test) Consequently wefelt confident that the tests of hypotheses could proceed in that the second-order variables were meaningfully representing the common attribute underlying allvariables within a set

As observed in the bottom third of Table 3 all of the hypothesized associationsamong the second-order variables were completely supported Specifically thesecond-order organizational attribute component had statistically significant andpositive associations with the second-order component of organizational climate(r2 = 95) which had significant positive associations with the second order com-ponents of job design (r2 = 78) and job future (r2 = 93) Hence as organizationsstrengthen their attributes (ie their policies regarding employee involvement safetyand health values concerning people etc) they can expect to see a correspondingstrengthening of perceptions regarding the climates in the organization (ie for com-munication involvement etc) In turn as employeesrsquo perceptions regarding theclimates within the organization are raised there is an associated increase regardingthe way people relate to the job (ie the degree of autonomy etc) and their future inthe organization (ie job security opportunities to learn new things etc)

Similar support was observed for the relationships from job design and job future towork adjustment (r2 = 83) Therefore a strengthening in job design (ie increasing itspositive attributes such as autonomy and decreasing its negative attributes such asworkload) is associated with a strengthening in psychological work adjustment (ieincreases in its positive attributes such as job satisfaction and decreases in its negativeattributes such as job stress) Similarly a strengthening in job future (ie increasing jobsecurity learning opportunities etc) would be associated also with a strengthening inpsychological work adjustment

Additionally as we strengthen psychological work adjustment (ie increase itspositive attributes such as job satisfaction and decrease its negative attributes such asjob stress) corresponding decreases in alcohol consumption (r2 = 01) tobacco use(r2 = 01) intentions to quit (r2 = 24) and absenteeism (r2 = 04) and increases inpsychological health (r2 = 17) and perceptions of general health (r2 = 02) are evidentIndeed an examination of the unstandardized coefficient for alcohol consumption(minus47) indicates that for every unit of strengthening in work adjustment employeesconsume approximately five less drinks per month That is not to imply that there is acause-effect relationship between work adjustment and alcohol consumption just thatthere is a strong relationship between the two In a similar vein tobacco use gets

580 Mark G Wilson et al

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 17: Work Characteristics and Employee Health

Healthy work organization 581

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 18: Work Characteristics and Employee Health

closer to zero (the unstandardized coefficient was minus04) with a corresponding unitstrengthening in work adjustment

Discussion

The results are consistent with and provide support for the proposed healthy organi-zation model Clearly work characteristics influence psychological work adjustmentfactors that ultimately affect employee health and well-being Although considerableprevious research has supported the association between certain work characteristicsand various measures of health and well-being this research is the first to provideempirical support for a comprehensive model that explains the nature of the relation-ships among these highly researched constructs Owing to the cross-sectional design ofthe study and the complexity of the interrelationships among the proposed variables acausendasheffect relationship should not be inferred from these findings However this isan important first step to understanding how organizational actions impact employeesThe proposed model demonstrated a good fit overall as well as significant associationsamong all of the hypothesized second-order latent variables Rarely does one see acomplex model of this nature where all hypothesized associations among second-orderconstructs and second- to first-order constructs demonstrate statistical significance

Even though the cross-sectional design restricts conclusions about causality thestudy clearly supports the crucial role that employees and their perceptions andexpectations play in organizational outcomes particularly health and well-beingoutcomes Policies procedures and actions that an organization takes to improveefficiency and effectiveness are filtered through the employees and reflected in theirsatisfaction with and commitment to the job and organization Hence any efforts toimpact organizational effectiveness must include strategies targeted toward employeesto at a minimum facilitate their understanding and buy-in Strategies that involve opencommunication (DeJoy et al 1995 McAfee amp Winn 1989 Schurman amp Israel 1995)and broad-based participation (Vandenberg et al 1999) have been shown to beimportant Considerable interest has been generated around the use of high-involvement work processes as a means of fostering employee involvement inorganizational initiatives (Eastman amp Vandenberg 1998 Edwards amp Wright 2001Lawler 1992 Riordan Vandenberg amp Richardson 2003 Vandenberg et al 1999)

This study also supports the fundamental role organizational climate plays in theeffectiveness of an organization The dimensions included in the climate domain are allrelated to providing support for employees either directly such as through coworkersupport and participation with others and supervisors or indirectly through a support-ive environment Clearly this reinforces the importance of the interpersonal aspects ofwork and emphasizes how much they influence the nature of work and the employ-ment relationship This social domain of work is probably the most intriguing and leastunderstood of the constructs studied and yet has a major influence on the efficiencyand effectiveness of the organization

One possible limitation of the present study is that all data were collected from thesame respondents using the same instrumentndashmaking common method variance apotential problem However we feel that this issue was not a significant factor for tworeasons First reviewing the correlations among first-order constructs in Table 1 indi-cates that across (ie as opposed to within) second-order constructs all first-orderconstructs were only weakly to moderately correlated This level of correlation sug-gests that inflation due to common method variance is not a problem In other words

582 Mark G Wilson et al

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 19: Work Characteristics and Employee Health

if substantial common method variance was present one would expect many if notmost of the constructs to be very highly correlated (ie |80| or higher) Secondrecent empirical research has indicated that this inflation due to common methodvariance is not as widespread as initially thought (Crampton amp Wagner 1994) In factmany of the variables that Crampton and Wagner cite in their analysis are similar to thevariables used in this study (eg pay and benefits career advancement organizationalculture job involvement)

A second possible limitation was related to the level of analysis This model test wasconducted at the individual level However the underlying conceptual premises of themodel could suggest the need to examine the model at the organizational level Forexample historically the term lsquoclimatersquo denotes a higher level in that it conveys a sensethat each unit varies with respect to the degree in which it promotes stronger com-munication participation and other elements within that particular unit suggestingthat the organizational level of analysis would be most appropriate At the same timeother dimensions such as those in the job-design domain (job content autonomyetc) are clearly conceptualized at the individual level as were many of the outcomevariables (ie depression anger alcohol use absenteeism) Since it was not possibleto conduct the analysis at both levels the data were collected from one organizationand the variables were measured at the individual level it made the most sense toconduct the analysis at the individual level Future research needs to examine theseconstructs across multiple organizations andor address this conceptual quandary

Finally these findings have implications for both research and practice Forresearchers this model provides an integrated conceptual framework from which tostudy healthy work organization The subject of organization of work and health hasyet to become a cohesive field of study (NIOSH 2002) As indicated in our literaturereview multiple disciplines have contributed to the current state of knowledge on thevarious dimensions of healthy work organization yet little cross-fertilization existsamong the disciplines This proposed model was an endeavour to draw the knowledgebase of work organization together and to provide an initial test of a comprehensivemodel An interdisciplinary approach to the study of healthy work organization canresult in the leaps of knowledge that are necessary to move this field of study forwardand keep pace with the constantly changing conditions of work in the private sector

Considerable work lies ahead in validating the measures and establishing causendasheffect relationships among the proposed first- and second-order latent variablesAdditional testing and model refinement are an important next step to provide theconfidence necessary for extended organizational application The model should alsobe tested in various settings with other organizational outcomes of interest Forexample health-care organizations are concerned with nurse retention and vacancyrates while public-sector organizations look at the provision of services as a measure ofeffectiveness In addition the model needs to be tested using financial variables as ameasure of organizational effectiveness Demonstrating a relationship between thehealthy work organization dimensions and the financial success of the organizationwould provide a strong argument to business for investing in healthier workplaces

For managers the findings indicate that organizational action is central to creating ormaintaining a healthy work organization Policies and procedures reflective of theorganizationrsquos values and beliefs serve as the foundation for organizational changeAt the same time multiple factors must be impacted to effect organizationalchange Simply modifying the work environment or increasing employee pay will notof and by themselves create the genesis necessary to significantly impact organizational

Healthy work organization 583

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 20: Work Characteristics and Employee Health

effectiveness The proposed model can provide a foundation for intervention as wellhelping organizations identify problem areas and providing a framework for tacklingtough issues Ideally interventions should foster broad-based participation on the partof employees throughout the organization encouraging open communication empow-ering employees to take action to make a difference in their organization and recogni-zing and rewarding them for their efforts

Acknowledgements

This work was supported in part by the National Institute for Occupational Safety and Health(NIOSH) and the US Centers for Disease Control and Prevention (CDC) (Grant [5-R01-OH03737-02) However its contents are solely the responsibility of the authors and do not necessarilyrepresent the official views of NIOSH or CDC

References

Arbuckle J L amp Wothke W (1999) Amos 40 Userrsquos Guide Chicago IL SmallWatersCorporation

Arthur J B (1994) Effects of human resource systems on manufacturing performance andturnover Academy of Management Journal 37 670ndash687

Bavendam J M Boyer M amp Sorensen W B (1986) Distributive justice In JL Price amp CWMueller (Eds) Handbook of organizational measurement Marshfield MA Pitman

Bohen H amp Viveros-Long A (1981) Balancing jobs and family life Philadelphia TempleUniversity Press

Caplan R D Cobb S French T R P Harrison R V amp Pinneau S R (1975) Job demandsand worker health Washington DC US Government Printing Office

Cartwright S Cooper C L amp Murphy L R (1995) Diagnosing a healthy organization Aproactive approach to stress in the workplace In L R Murphy J J Hurrell Jr S L Sauter ampG P Keita (Eds) Job stress interventions (pp 217ndash233) Washington DC AmericanPsychological Association

Cheyne A Cox S Oliver A amp Tomas J M (1998) Modeling safety climate in the predictionof levels of safety activity Work amp Stress 12 255ndash271

Cohen S Kamarck T amp Mermelstein R (1983) A global measure or perceived stress Journalof Health and Social Behavior 24 385ndash396

Cooper C L amp Cartwright S (1994) Healthy mind Healthy organizationmdashA proactiveapproach to occupational stress Human Relations 47 455ndash471

Cooper C L amp Williams S (1994) Creating healthy work organizations Chichester UKWiley

Cox T (1988) Editorial Organizational health Work amp Stress 2 1ndash2Cox T amp Howarth I (1990) Organizational health culture and helping Work amp Stress 4

107ndash110Cox T Leather P amp Cox S (1990) Stress health and organisations Occupational Health

Review FebruaryMarch 13ndash18Crampton S M amp Wagner J A (1994) Perceptndashpercept inflation in microorganizational

research An investigation of prevalence and effect Journal of Applied Psychology 79 67ndash76Danna K amp Griffin R W (1999) Health and well-being in the workplace A review and

synthesis of the literature Journal of Management 25 357ndash384DeJoy D M Murphy L amp Gershon R M (1995) The influence of employee jobtask and

organizational factors on adherence to universal precautions among nurses InternationalJournal of Industrial Ergonomics 16 43ndash55

DeJoy D M Searcy C A Murphy L R amp Gershon R R M (2000) Behavioral-diagnosticanalysis of compliance with universal precautions among nurses Journal of OccupationalHealth Psychology 5 127ndash141

584 Mark G Wilson et al

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 21: Work Characteristics and Employee Health

DeJoy D M amp Southern D J (1993) An integrative perspective on worksite health promotionJournal of Occupational Medicine 35 1221ndash1230

DeJoy D M amp Wilson M G (2003) Organizational health promotion Broadening the horizonof workplace health promotion American Journal of Health Promotion 17 337ndash341

Delery J E amp Shaw J D (2001) The strategic management of people in work organizationsReview synthesis and extension In G R Ferris (Ed) Research in personnel and humanresources management Greenwich CT Elsevier

Demerouti E Bakker A B Nachreiner F amp Schaufeli W B (2001) The job demands-resources model of burnout Journal of Applied Psychology 86 499ndash512

Du Toit M amp Du Toit S (2001) Interactive LISREL Userrsquos Guide Lincolnwood IL ScientificSoftware International

Eastman L J amp Vandenberg R J (1998) Cultures for excellence A research reportlinking quality of work life quality service and employee involvement with competitiveadvantage Atlanta GA Life Office Management Association

Edwards P amp Wright M (2001) High-involvement work systems and performance outcomesThe strength of variable contingent and context-bound relationships International Journalof Human Resource Management 12 568ndash585

Eisenberger R Huntington R Hutchison S amp Sowa D (1986) Perceived organizationalsupport Journal of Applied Psychology 71 500ndash507

Ferris G R Hochwarter W A Buckley M R Harrell-Cook G amp Frink D D (1999) Humanresources management Some new directions Journal of Management 25 385ndash416

Gerbing D W amp Anderson J C (1987) Improper solutions in the analysis of covariancestructures Their interpretability and a comparison of alternative specificationsPsychometrika 52 99ndash111

Goetzel R Z Jacobson B H Aldana S G Vardell K amp Yee L (2000) Health care costs ofworksite health promotion participants and non-participants Journal of Occupational andEnvironmental Medicine 40 341ndash346

Greenberg J (1986) Determinants of perceived fairness of performance evaluations Journal ofApplied Psychology 71 340ndash342

Griffin M A amp Neal A (2000) Perceptions of safety at work A framework for linking safetyclimate to safety performance knowledge and motivation Journal of Occupational HealthPsychology 5 347ndash358

Hackman J R amp Oldham G R (1975) Development of the job diagnostic survey Journal ofApplied Psychology 60 159ndash170

Hackman J R amp Oldham G R (1980) Work redesign Reading MA Addison-WesleyHofmann D A Jacobs R amp Landy F (1995) High reliability process industries Individual

micro and macro organizational influences on safety performance Journal of SafetyResearch 26 131ndash149

House J S McMichael A J Wells J A Kaplan B H amp Landerman L R (1979)Occupational stress and health among factory workers Journal of Health and SocialBehavior 20 139ndash160

Huselid M A (1995) The impact of human resource management practices on turnoverproductivity and corporate financial performance Academy of Management Journal 38635ndash672

Ilfeld F W (1978) Psychologic status of community residents along major demographicdimensions Archives of General Psychiatry 35 716ndash724

Jaffe D T (1995) The healthy company Research paradigms for personal and organizationalhealth In S L Sauter amp L R Murphy (Eds) Organizational risk factors for job stress(pp 13ndash39) Washington DC Amercian Psychological Association

James L R amp McIntyre M D (1996) Perceptions of organizational climate In Murphy K(Ed) Individual differences and behavior in organizations San Francisco Jossey-Bass

Jamieson D amp OrsquoMarra J (1991) Managing workforce 2000 Gaining the diversity advan-tage San Francisco CA Jossey-Bass

Healthy work organization 585

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 22: Work Characteristics and Employee Health

Johansson G Johnson J V amp Hall E M (1991) Smoking and sedentary behavior as related towork organization Social Science and Medicine 32 837ndash846

Klitzman S House J S Israel B A amp Mero R P (1990) Work stress nonwork stress andhealth Journal of Behavioral Medicine 13 221ndash243

Kuhnert K Sims R R amp Lahey M A (1989) The relationship between job security andemployee health Group amp Organization Studies 14 399ndash410

Landsbergis P A (2003) The changing organization of work and the safety and health ofworking people A commentary Journal of Occupational and Environmental Medicine45 61ndash72

Lawler E E III (1992) The ultimate advantage Creating the high involvement organizationSan Francisco Jossey-Bass

Lindstrom K (1994) Psychosocial criteria for good work organization Scandanavian Journalof Work Environmental Health 20 123ndash133

Lowe G S Schellenberg G amp Shannon H S (2003) Correlates of employeesrsquo perceptions ofa healthy work environment American Journal of Health Promotion 17 390ndash399

McAfee R B amp Winn A R (1989) The use of incentivesfeedback to enhance workplacesafety A critique of the literature Journal of Safety Research 20 7ndash19

McDonald R P amp Marsh H W (1990) Choosing a multivariate model Noncentrality andgoodness of fit Psychological Bulletin 107 247ndash255

Morrow P C McElroy J C amp Elliot S M (1994) The effect of preference for work statusschedule and shift on work-related attitudes Journal of Vocational Behavior 45 202ndash222

Mowday R T Steers R M amp Porter L W (1979) The measurement of organizationalcommitment Journal of Vocational Behavior 14 224ndash247

National Institute for Occupational Safety and Health (1996) National occupational researchagenda Washington DC Author

National Institute of Occupational Safety and Health (2002) The changing organization ofwork and the safety and health of working people Washington DC Author

OrsquoReilly C A Chatman J amp Caldwell D F (1991) People and organizational culture A profilecomparison approach to assessing personndashorganization fit Academy of ManagementJournal 34 487ndash516

Ostroff C amp Bowen D E (2000) Moving HR to a higher level HR practices and organizationaleffectiveness In K J Klein amp S W J Kozlowski (Eds) Multilevel theory research andmethods in organizations foundations extensions and new directions San FranciscoJossey-Bass

Ozminkowski R J Dunn R L Goetzel R Z Cantor R I Murnane J amp Harrison M (1999)A return on investment evaluation of the Citibank NA health management programAmerican Journal of Health Promotion 14 31ndash43

Parker S amp Wall T (1998) Job and work design Organizing work to promote well-beingand effectiveness Thousand Oaks CA Sage

Pate-Cornell M E (1990) Organizational aspects of engineering system safety The case ofoffshore platforms Science 250 1210ndash1217

Pelletier K R (1984) Healthy people in unhealthy places stress and fitness at work NewYork Delacorte Press

Peterson M amp Wilson J F (2002) The culturendashworkndashhealth model and work stress AmericanJournal of Health Behavior 26 16ndash24

Reason J (1990) Human error New York Cambridge University PressReason J (1995) A systems approach to organizational error Ergonomics 38 1708ndash1721Ribisl K M amp Reischl T M (1993) Measuring the climate for health at organizations

Development of the worksite health climate scales Journal of Occupational Medicine 35812ndash824

Riordan C M Vandenberg R J amp Richardson H A (2003) Employee involvement andorganizational effectiveness An organizational systems perspective Manuscript submittedfor publication

586 Mark G Wilson et al

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 23: Work Characteristics and Employee Health

Rizzo J R House R J amp Lirtzman S I (1970) Role conflict and ambiguity in complexorganizations Administrative Science Quarterly 206 150ndash163

Rokeach M (1979) From individual to institutional values With special reference to the valuesof science In M Rokeach (Ed) Understanding human values (pp 47ndash70) New York FreePress

Rosen R H amp Berger L (1991) The healthy company Eight strategies to develop peopleproductivity and profits New York GP Putnamrsquos Sons

Rousseau D M (1997) Organizational behavior in the new organizational era Annual Reviewof Psychology 48 515ndash546

Sainfort F Karsh B T Booske B C amp Smith M J (2001) Applying quality improvementprinciples to achieve healthy work organizations Journal on Quality Improvement 27469ndash483

Sandroff D J Bradford S amp Gilligan V F (1990) Meeting the health promotion challengethrough a model of shared responsibility Occupational Medicine State of the Art Review 5677ndash690

Sauter S L Lim S amp Murphy L R (1996) Organizational health A new paradigm foroccupational stress research at NIOSH Occupational Mental Health 4 248ndash254

Sauter S L Murphy L R amp Hurrell J J Jr (1990) Prevention of work-related psychologicaldisorders A national strategy proposed by the National Institute for Occupational Safety andHealth (NIOSH) American Psychologist 45 1146ndash1158

Sawyer J (1992) Goal and process clarity Specification of multiple constructs of role ambiguityand a structural equation model of their antecedents and consequences Journal of AppliedPsychology 77 130ndash142

Schein E H (1990) Organizational culture American Psychologist 45 109ndash119Schurman S J amp Israel B A (1995) Redesigning work systems to reduce stress A participatory

action research approach to creating change In L R Murphy J J Hurrell S L Sauter ampG P Keita (Eds) Job stress interventions Washington DC American PsychologicalAssociation

Shannon H S Robson L S amp Sale J E M (2001) Creating safer and healthier workplacesRole of organizational factors and job characteristics American Journal of IndustrialMedicine 40 319ndash334

Smith K K Kaminstein D S amp Makadok R J (1995) The health of the corporate bodyIllness and organizational dynamics Journal of Applied Behavioral Science 31 328ndash351

Sparks K Faragher B amp Cooper C L (2001) Well-being and occupational health in the 21stcentury workplace Journal of Occupational and Organizational Psychology 74 489ndash509

Spreitzer G M (1995) Psychological empowerment in the workplace Dimensions measure-ment and validation Academy of Management Journal 38 1442ndash1465

Steiger J H (1990) Structural model evaluation and modification An interval estimationapproach Multivariate Behavioral Research 25 173ndash180

Stokols D (1992) Establishing and maintaining healthy environments Toward a social ecologyof health promotion American Psychologist 47 6ndash22

Tucker L R amp Lewis C (1973) A reliability coefficient for maximum likelihood factor analysisPsychometrika 38 1ndash10

Vandenberg R J Park K DeJoy D M Wilson M G amp Griffin-Blake C S (2002) Thehealthy work organization model Expanding the view of individual health and well being inthe workplace In P L Perrewe amp D C Ganster (Eds) Historical and current perspectiveson stress and health Oxford Elsevier

Vandenberg R J Richardson H amp Eastman L (1999) High involvement organizations Theirantecedents and consequences Groups amp Organizations Management 24 300ndash339

Vroom V H (1959) Some personality determinants of the effects of participation Journal ofAbnormal and Social Psychology 59 322ndash327

Ware J E amp Sherbourne C D (1992) The MOS 36-item short-form health survey (SF-36)Medical Care 30 473ndash481

Healthy work organization 587

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al

Page 24: Work Characteristics and Employee Health

Warr P (1994) A conceptual framework for the study of work and mental health Work ampStress 8 84ndash97

Received 3 March 2003 revised version received 25 November 2003

588 Mark G Wilson et al