Development and Field Evaluation of an Interdisciplinary ... · job enrichment and motivating job...

Journal of Applied Psychology1985, Vol. 70, No. 1, 29-43

Copyright 1985 by the American Psychological Association, Inc.0021-90IO/85/S00.75

Development and Field Evaluation of anInterdisciplinary Measure of Job Design

Michael A. CampionInternational Business Machines Corporation

Research Triangle Park, North Carolina

Paul W. ThayerNorth Carolina State University

The development of an interdisciplinary job design questionnaire and a study ofits interrelationships with a variety of outcomes is described. A taxonomy of jobdesign approaches was developed from literature of different disciplines: (a) amotivational approach from organizational psychology; (b) a mechanistic approachfrom classic industrial engineering; (c) a biological approach from work physiologyand biomechanics; and (d) a perceptual/motor approach from experimentalpsychology. The Multimethod Job Design Questionnaire (MJDQ) was developedreflecting these approaches. A corresponding taxonomy of job outcomes wasdeveloped, and hypotheses were generated as to relationships between job designapproaches and outcomes. A field study involved 121 jobs, 215 incumbents, and23 supervisors from five plants. Results indicated the MJDQ was reliable, andmost hypotheses were supported. Different job design approaches influencedifferent outcomes and may have some costs as well as benefits; an interdisciplinaryperspective is needed to integrate major theories of job design.

Even a cursory examination of the jobdesign literature reveals many different schoolsof thought: industrial engineering approachesof scientific management and time and mo-tion study, the psychological approaches ofjob enrichment and motivating job charac-teristics, the human factors or ergonomicsapproaches, and sociotechnical approachesto job design. Although there is some overlapin the recommendations made for proper jobdesign, there is considerable divergence infocus and even some direct conflict in advice.Proponents, however, claim that their jobdesigns positively influence most of the out-come spectrum for both the individual andthe organization.

This study won the 1983 S. Rains Wallace DissertationAward sponsored by the Society of Industrial/Organiza-tional Psychology, Division 14 of the American Psycho-logical Association. It was conducted while the firstauthor was a doctoral student at North Carolina StateUniversity.

Special thanks to Kitty Klein, Richard Pearson, andMichael Joost for their comments and suggestions onthis research. Thanks also to the many managers andemployees of Weyerhaeuser Company who contributedtime and data to this study.

Requests for reprints should be sent to Michael A.Campion, IBM Corporation, D673/B205, P.O. Box 12195,Research Triangle Park, North Carolina 27709.

The present study addresses this confusionby pulling together the diverse literature onjob design, delineating major approaches,and demonstrating that each approach isgeared toward a particular subset of outcomes.More specifically, this study (a) develops ajob design taxonomy, (b) develops a corre-sponding job outcome taxonomy, (c) developsmeasures that reflect the design taxonomy,(d) develops measures which reflect the out-come taxonomy, and (e) evaluates differentialpredictions of job design-outcome relation-ships in a field setting.

Taxonomy of Job Design Approaches

The first step was to consult the literatureand extract specific job design rules. Nearly700 job design rules resulted, suggesting ad-equate coverage of the content domain. Ruleswere then sorted into fairly homogeneousgroups based on underlying theoretical per-spectives. Similar rules were combined intoa principle that summarized their main con-tent. Principles were written to represent theconsensus from the literature, each reflectingcommon content from a large number ofspecific rules. They were also broad enoughto be applicable across diverse jobs, yet spe-cific enough to allow objective and quantifi-

29

30 MICHAEL A. CAMPION AND PAUL W. THAYER

able judgments about jobs. The end productwas four sets of principles, each set consti-tuting a job design approach.

The content coverage of the resulting setsof principles appeared adequate because ofthe variety of literature consulted, the numberof job design rules discovered, the fact thatonly two percent of the rules could not begrouped into design approaches, and the factthat only 11% of the rules could not becombined into the job design principles. Re-producibility was assessed by having a naivejudge reclassify the principles back into jobdesign approaches, resulting in 83.8% agree-ment.

The four job design approaches thatemerged are described below.

(Barnes, 1980; Konz, 1979; Mundel, 1970;Nadler, 1963). The 13 principles are in Table2. The discipline base is classic industrialengineering. "Classic" is used because manymodern day writers on industrial engineeringaddress a variety of job design approaches(e.g., Konz, 1979).

Biological

This approach derives from a book onbiomechanics (Tichauer, 1978), articles onposture (Ayoub, 1973; Floyd & Ward, 1966;Grandjean & Hunting, 1977; Van Wely, 1970)and lifting strength (Chaffin, 1974; Park &Chaffin, 1975; Snook & Irvine, 1967), books

Motivational

This approach came from literature on jobenrichment and enlargement (Ford, 1969;Herzberg, 1966, 1968; Hulin and Blood, 1968;Mayer, 1971; Walters, 1975), research andreviews on characteristics of motivating jobs(Griffin, 1982; Hackman & Lawler, 1971;Hackman & Oldham, 1980; Pierce & Dun-ham, 1976; Turner & Lawrence, 1965), in-struments used to measure jobs' motivatingfeatures (Hackman & Oldham, 1975; Jenkins,Nadler, Lawler, & Cammann, 1975; Sims,Szilagyi, & Keller, 1976), theories of workmotivation and organizational behavior (Ar-gyris, 1964; Likert, 1961; McGregor, I960;Mitchell, 1976; Steers & Mowday, 1977;Vroom, 1964), texts in industrial and orga-nizational psychology (Cascio, 1978; Dun-nette, 1976; Wexley & Yukl, 1977), andpsychological principles from sociotechnicaldesign approaches (Cherns, 1976; Engelstad,1979; Rousseau, 1977). The 16 principlesextracted are contained in Table 1. The maindiscipline base is organizational psychology.

Mechanistic

These principles were extracted from classictexts on scientific management (F. Taylor,1911) and motion study (Gilbreth, 1911),two main handbooks of industrial engineering(Ireson & Grant, 1971; Maynard, 1971), andtexts by other writers on time and motionstudy, work simplification, and specialization

Table 1Descriptions, Means, Standard Deviations, andItem-Total Correlations on the Items in theMotivational Job Design Scale

Description M SD

1. Autonomy,responsibility,verticalloading

2. Intrinsic jobfeedback

3. Extrinsic jobfeedback

4. Socialinteraction

5. Task/goal clarity6. Task variety,

horizontalloading

7. Ability/skillrequirementsand variety

8. Task identity9. Task significance

10. Growth,learning,advancingresponsibility

11. Promotion12. Achievement13. Participation14. Communication15. Pay adequacy16. Job security

Total score"

121

121

121

121121

121

121121121

121121121121121121121121

3.16

3.79

3.22

3.164.06

2.84

2.643.313.07

2.953.073.052.862.893.203.053.15

.94

.72

.82

.76

.47

.96

.87

.94

.96

.90

.94

.70

.85

.67

.40

.67

.48

.74

.43

.76

.38

.19

.56

.82

.71

.81

.67-.69

.73

.83

.68

.18

.68

.89C

'Item-total correlations. AH correlations significant atp < .05.b Based on average of all items per job.c Coefficient alpha (Cronbach, 1951).

INTERDISCIPLINARY MEASURE OF JOB DESIGN 31

on work physiology (Astrand & Rodahl, 1977)and anthropometry (Hertzberg, 1972; Roe-buck, Kroemer, & Thompson, 1975), ergo-nomie texts that cover many approaches in-cluding the biological approach (e.g., Grand-jean, 1980), and industrial engineering texts,which include sections on biological ap-proaches (H. Davis & Miller, 1971; Konz,1979). The 18 principles composing this ap-proach are contained in Table 3. The disci-pline bases are the biological sciences, espe-

Table 2Descriptions, Means, Standard Deviations, andItem-Total Correlations on the Items in theMechanistic Job Design Scale

Description n M SD r*

1. Taskfractionalization/specialization 121 3.77 1.02 ,73

2. Specialization ofmaterials, tools,procedures 121 3.98 .77 .52

3. Task simplification 121 3.82 1.01 .504. Skill simplification 121 3.60 1.00 .475. Repetition/pacing 121 3.67 .96 .606. Idle time/capacity 121 3.74 .96 .267. Motion

economy —materialshandling 78 3.88 .79 .55

8. Motion

Table 3Descriptions, Means, Standard Deviations, andItem-Total Correlations on the Items in theBiological Job Design Scale

Description n M SD r'

1. Seating 121 3.02 1.14 .692. Tool design 15 3.33 .49 .923. Anthropometry 121 3.29 .68 .574. Static effort 121 3.47 .91 .675. Endurance 121 3.26 .85 .756. Strength 121 3.28 .78 .537. Lifting 121 3.31 .88 .568. Posture, lower back 121 2.92 .98 .779. Muscular adequacy 119 3.60 .73 .71

10. Wrists 121 2.95 1.09 .5811. Stress concentration 121 3.20 .70 .7012. Vibration 121 3.11 1.05 .1913. Noise 121 2.69 .89 .5914. Climate 121 1.70 .82 .5515. Atmosphere 121 2.90 .92 .3016. Worker

protection —safety 121 4.15 1.10 .49

17. Work breaks 121 3.40 .69 .3118. Shiftwork 121 3.85 .56 -.15

Total score" 121 3.18 .52 .86°

* Item-total correlations. Items with missing data estimatedas mean of applicable items for each job and included inthe calculations. All correlations significant at p < .05.b Based on average of applicable items per job.c Coefficient alpha (Cronbach, 1951).

cially work physiology, biomechanics, andanthropometry.

economy—pre-positioning ofmaterials, tools

9. Motioneconomy—eyeand headmovements

10. Motioneconomy—musclemovement

11. Motioneconomy—muscle rhythm

12. Motioneconomy—musclemovement style

13. MechanizationTotal score"

70 4.00 .72

106 2.81 1.03

107 3.28 .84

73 3.85 .49

74 3.66 .71117 2.86 .84121 3.55 .48

.78

.46

.29

.54

.45

.15

.82C

" Item-total correlations. Items with missing data estimatedas mean of applicable items for each job and included inthe calculations. All correlations significant at p < .05." Based on average of applicable items per job.c Coefficient alpha (Cronbach, 1951).

Perceptual/Motor

Relevant literature includes the manyhandbooks on human engineering (Mc-Cormick, 1976; Morgan, Cook, Chapanis, &Lund, 1963; Van Cott & Kinkade, 1972;Woodson, 1954, 1981; Woodson & Conover,1964), a text that deals with many aspects ofhuman factors or ergonomics (Grandjean,1980), literature on skilled performance(Kahneman, 1973; Welford, 1976), and theo-retical treatments of people as informationprocessors (Fogel, 1967; Gagne, 1962; Good-stein, 1981; Rasmussen, 1981). The 23 prin-ciples are in Table 4. The base is experimentalpsychology.

Taxonomy of Job Outcomes

A taxonomy of job outcomes was developedfrom the literature to correspond to the tax-


onomy of job design approaches. An exami-nation of the content of each approach, thedependent variables used in research, and theunderlying theories revealed that each ap-proach was actually geared toward a specificcategory of outcomes. Each outcome categorywas fairly cohesive and homogeneous, repre-senting a common theme or purpose. Thefour categories of job outcomes are describedbelow with illustrative measures of each.

Table 4Descriptions, Means, Standard Deviations, andItem-Total Correlations on the Items in thePerceptual/Motor Job Design Scale

Description n M SD

1.

2.

3.

4.5.

6.

7.8.

9.

10.

11.

12.13.14.15.16.17.

18.19.20.21.22.23.

Workplace lighting —general

Workplace lighting — glare/contrast

Control and displayidentification

Display visibility/legibilityDisplays — information

contentControl/display movement

relationshipsControl/display ratiosControl resistance/

feedbackControls — accidental

activationControls — anthropometry/

biomechanicsControls — motion

economyWarning devicesPrinted job materialsPanel layoutInput requirementsOutput requirementsInformation processing

requirementsMemory requirementsBoredomArousalStressWorkplace layout — safetyWorkplace layout — visual

and auditory linksTotal score11

121

121

8442

42

8175

79

85

82

83112976

121121

121121121121121121

121121

3.22

3.22

3.153.64

4.05

3.523.71

3.54

2.54

3.40

3.803.823.903.283.383.36

3.663.853.003.163.052.73

3.223.36

.57

.58

.90

.58

.31

.62

.56

.57

.89

.78

.60

.41

.31

.74

.76

.78

.82

.91

.85

.83

.93

.95

.65

.36

.37

.33

.56

.60

.62

.54

.42

.54

.48

.53

.35

.81

.71

.46

.56

.63

.46

.34-.05

.37

.58

.40

.65

.85C

* Item-total correlations. Items with missing data estimatedas mean of applicable items for each job and included inthe calculations. All correlations, except for item 19, sig-nificant at p < .05.b Based on average of applicable items per job.' Coefficient alpha (Cronbaeh, 1951).

Satisfaction

This category refers to affective, motiva-tional, or attitudinal outcomes from worksuch as job satisfaction, job involvement, andintrinsic work motivation. It includes certainbehavioral indexes such as job performanceand absenteeism.

Efficiency

This category refers to human resourceefficiency and flexibility. Measures includeestimates of utilization levels or the percentageof people who could perform the job. Othermeasures include estimates of training timeand measures estimating idle time on the job.

Comfort

This category includes most of the physicalwell-being outcomes. Measures include sub-jective effort, physical fatigue, comfort, healthrecords, and health complaints, such as re-ports of back problems, muscle strain, andhearing loss.

Reliability

This category may seem less obvious thanthe others. It refers to safety, system reliability,and user reactions to equipment, facilities, orworkplaces. System safety variables such asaccident rates and accident-prone situations(Swain, 1973), as well as medical data oninjuries, are included. System reliability isreflected in measures of error rate or error-likely situations (Swain, 1973). Another com-mon theme is user reaction toward the systemas to work overload or underload, mentalfatigue or task aversion (Hartley and Chute,1947), and attitudes toward equipment.

Hypotheses

The motivational job design approach ispredicted to correlate most positively withthe satisfaction outcome category, as it wasoriginally derived from that literature. On aspeculative basis, the motivational approachmay correlate negatively with the efficiencyoutcome category, because more motivatingand satisfying jobs often involve higher skilllevels and more responsibility and thus wouldexhibit lower utilization levels and increased


training time. The motivational approach isnot logically expected to correlate with theother two outcome categories.

The mechanistic job design approach ispredicted to correlate most positively withthe efficiency outcome category, because theprimary goals of classic industrial engineeringpractices were economic. It is also speculatedthat the mechanistic approach may correlatepositively with the reliability category due toits concern with safety and reliability. Themechanistic approach is expected to correlatenegatively with the satisfaction and comfortcategories because of the overwhelming evi-dence on the negative attitudinal and healthconsequences of mechanistically designedwork (e.g., Caplan, Cobb, French, Van Har-rison, & Pinneau, 1975; Frankenhaeuser,1977; Johansson, Aronsson, & Lindstrom,1978; Karasek, 1979; Kornhauser, 1965; Sal-vendy & Smith, 1981; Shepard, 1969, 1970;Walker & Guest, 1952; Weber, Fussier,O'Hanlon, Gierer, & Grandjean, 1980).

The biological job design approach isclearly most concerned with the physicalwell-being of the worker and thus is predictedto correlate most positively with the comfortoutcome category. The approach may also bepositively correlated with all the other cate-gories because healthier and more comfortableworkers may be more satisfied, efficient, andreliable.

The perceptual/motor approach is pre-dicted to correlate most positively with thereliability outcome category. The primarygoal is to enhance the person-machine fit byattending to people's perceptual/motor ca-pabilities and limitations, thus preventingerrors and accidents and reducing boredomand task aversion. Positive user regard isanother indication that this match has beenachieved. Positive, or at least nonnegative,correlations will exist with the other categoriesbecause good person-machine fit might en-hance other attitudinal outcomes, efficienthuman resource utilization, and effort orcomfort.

Support for convergent and discriminantvalidity (Campbell & Fiske, 1959) of themain hypotheses will be indicated if each jobdesign approach is more positively correlatedwith that one outcome category than theothers. Support for directional, speculative

hypotheses will come from relationships inthe predicted direction or near zero relation-ships; relationships in the direction oppositeto that predicted will indicate lack of support.

Method

Sample

Power analysis (Cohen, 1977) suggested a sample of112 jobs to detect a minimum correlation (between jobdesign scales and outcomes) of .30 with desired power of.90. The actual sample of jobs was 121, representing acomplete census of hourly production jobs in five Southernwood products operations of a large company: a plywoodplant (40 jobs), a sawmill (33 jobs), a fiberwood plant(25 jobs), a wood treatment (e.g., landscape timber)facility (14 jobs), and a merchandiser (i.e., log sorting/grading) facility (9 jobs). In terms of Equal EmploymentOpportunity (EEO) codes, the jobs were 23 skilled/craft,63 semiskilled/operative, 34 unskilled labor, and 1 clerical.Pay ranged from $5.25 to $9.18 per hour (M = 6.19,SD = .98).

Data were collected from two incumbents for 94 ofthe jobs and from one incumbent for the remainder.Thus, the sample of job incumbents was 215. They were77.2% male, 69.8% black, 29.3% white, and .9% Hispanic.Median age was 29.9 years (range = 19 to 63), medianeducation was 11.9 years (range = 2 to 15), mediancompany tenure was 5.7 years (range = .5 to 17), andmedian job tenure was 1.5 years (range = .1 to 12.5).

Data were also collected from two supervisors for 97of the jobs and from one supervisor for the remainder.Because each supervised many jobs, this resulted in datafrom 23 different supervisors. They were 87.0% male,30.4% black, and 69.6% white. Median age was 38.5years (range = 27 to 58), and median tenure as a super-visor was 9 years (range = 2 to 17).

Job Design Measurement

A number of methodological issues contributed tomeasurement decisions. First, an observational approachwas chosen because self-reports are susceptible to methodbias (Pierce & Dunham, 1976; Roberts & Click, 1981)and many forms of perceptual biases (Caldwell & O'Reilly,1982; O'Reilly & Caldwell, 1979; O'Reilly, Parlette, &Bloom, 1980; Shaw, 1980). Second, 5-point rating scaleswere anchored with verbal descriptions, definitions, orexamples to enhance interrater reliability. Anchor de-scriptions used adjectives whose psychophysical valuesaided discriminability (Bass, Cascio, & O'Connor, 1974).A 3-point relevance rating was also included for eachitem, and a space for comments permitted justificationof borderline ratings.

Because additive models are as good as or better thanmultiplicative models for combining job design elements(Pierce & Dunham, 1976), and unit weighting is generallypreferable to differential weighting schemes (Einhorn &Hogarth, 1975; Wainer, 1976), scores on job designprinciples were summed to form a composite withineach job design approach.

In summary, the job design analysis instrument is


applicable to a wide range of jobs. It has four sections,one for each approach, and is completed by an analystat the job site based primarily on observations. Scalepoints are anchored, and total scores for scales are simplesums.

An illustrative item from each of the four sections iscontained in the Appendix. The complete 70-item instru-ment is entitled the Multimethod Job Design Question-naire (MJDQ).1

In a pilot study, two experienced analysts completedMJDQs on 30 diverse jobs. Each job was observed for15 to 30 min with occasional informal questioning ofthe worker about less observable job aspects (e.g., infre-quent tasks). Questionnaires were then completed inde-pendently. Interrater reliabilities on the total scale scoresranged from .89 (p<.001) to .93. Mean agreementbetween raters across scales ranged from .12 to .17 onthe 5-point scale.2

Job Outcome Measurement

Multiple indexes were included for each outcomecategory, and data were collected from a variety ofsources. Interviews were used to collect much of the datain order to ensure thoughtful answers, minimize missingdata, and eliminate problems arising from the reading-level abilities of some respondents. Multiple sourcesincluded incumbents, supervisors, and archival records.An attempt was made to collect data from all threesources for each outcome category. Two incumbents andtwo supervisors were interviewed in most cases to reduceidiosyncratic biases. Questions were as objective as pos-sible, and descriptive anchors defined most rating scales.Finally, outcome measurement was guided by the literaturein each of the job design areas.3

Job incumbent interview. Pilot work indicated thatquestion complexity had to be appropriate to a widerange of educational backgrounds and communicable innoisy and distracting work sites. Questions could not beof a threatening nature (e.g., too efficiency oriented) ifhonest answers were desired.

A 23-item interview was developed. Six attitudinalitems assessed three of the constructs in the satisfactionoutcome category: two each on job satisfaction (Brayfield& Rothe, 1951), intrinsic work motivation (Hackman &Lawler, 1971), and job involvement (Lodahl & Kejner,1965). Four items assessed constructs in the efficiencycategory: two on estimates of utilization levels (i.e.,percentages of people who could perform the jobs), oneon training time, and one on idle time on the job. Fiveitems measured constructs in the comfort category: oneeach on physical effort (Borg, 1962) and physical fatigue(Kinsman & Weiser, 1976), two on various forms ofdiscomfort (e.g., backaches), and a checklist of healthcomplaints (Corlett & Bishop, 1976). Finally, eight itemsaddressed the reliability category: two on the accidentproneness of the job (Swain, 1973), one on error-likeli-hoods (Swain, 1973), three on work overload/underload(Ivancevich & Matteson, 1980; McCormick, 1976), oneon mental fatigue (Pearson, 1957), and one on attitudestoward equipment (Bare, 1966).

Supervisor interview. Information was collected fromsupervisors because they have observed many different

incumbents in each job, can better see interrelationshipsbetween jobs, and are not threatened by job evaluationquestions. Because each was to evaluate many jobs, theinterview was kept short by including only one or twoglobal questions for each outcome category. Many ques-tions took advantage of the supervisor's unique positionby asking for relative comparisons between jobs.

The supervisor interview contained 10 questions. Twowere checks to eliminate recently changed jobs andinexperienced supervisors. Neither was encountered. Rat-ings were included for each of the four outcome categories.One assessed job performance as part of the satisfactioncategory. The four items on efficiency asked of theincumbent were also in the supervisor interview. Oneoverall item on physical demands was included for thecomfort category. Finally, two scales, one on mentaldemands and another on error likelihoods, were includedfor the reliability category.

Archival data. Absenteeism is an archival measure inthe satisfaction category. Based on a review of problemswith this type of data (e.g., Hammer & Landau, 1981),information was collected on occurrences absent (Mdn -2.5 per 12 months, range = 0 to 8.9), days absent(Mdn = 4.2 per 12 months, range = 0 to 31.8), andpartially missed work days (Mdn =1.6 per 12 months,range = 0 to 5.4). It was not possible to distinguishvoluntary from involuntary absences due to recordkeepingdifferences among plants. Data were collected for theprevious 18 months and only on employees with at least6 months job experience (n = 169).

Medical data related to both the comfort and reliabilitycategories. Because of recordkeeping differences betweenplants, only the total number of medical incidents wasrecorded per employee for the entire job tenure, excludingthose with less than 6 months tenure (n = 169, Mdn =.5 per 12 months, range = 0 to 2.7).

Procedure

Measures were typically obtained in one departmentat a time until the plant was completed. The study wasexplained to managers and hourly employees throughproduction or safety meetings, followed by a tour of thedepartment with the supervisor. Next an MJDQ wascompleted on each hourly production job based primarilyon observation, with occasional informal questioning. Itwas completed before the outcome interviews to avoidexperimenter bias. Usually the two incumbents andsupervisors were from different shifts. To ensure thoughtfulanswers, considerable time was spent becoming familiarwith the respondents and explaining that the study wasnot company sponsored. Archival data were collectedfrom personnel and medical records just before theresearchers left the plant.

1 The MJDQ is available in M. A. Campion, M.A.(1985): The multimethod job design questionnaire(MJDQ). Psychological Documents, 15 (1) or from thefirst author.

2 Reliability and agreement analyses on individual itemsare available from the first author.

'Copies of all measurement protocols are availablefrom the first author.


Results

Analyses of Measures—Job Design

Tables 1 through 4 indicate that mostitems on the MJDQ are applicable to mostjobs. There is substantial variance; no rangerestriction is apparent. All but four of theitem-total correlations are positive and sig-nificant, and internal consistencies are in the.80s across the scales. Table 5 presents theintercorrelations among the job design scales.

Analysis of Measures—Outcomes

The 35 job outcomes showed reasonablevariance and no severe restriction.4 Two com-posites of the outcome items were formed toallow a simplified presentation of the results.First, an intercorrelated subset of items fromeach category of the outcome taxonomy wasformed into a composite via standardizedequal weighting. These composites are re-ferred to as the theoretical composites. Inter-nal consistencies are as follows: Satisfaction(5 items, alpha = .69), Efficiency (6 items,alpha = .72), Comfort (5 items, alpha = .66),and Reliability (8 items, alpha = .67). Inter-correlations are displayed in Table 6.

The second data-reduction approach wasto factor analyze the entire set of outcomeitems using varimax rotation. Five factors,explaining 78.3% of the variance, emergedwith eigenvalues exceeding 1.0. Factor scoreswere calculated and given descriptive labels.Their correlations with the theoretical com-posites are in Table 7. Notice that the MentalEase, Physical Ease, and Attitude Favorabilityfactors are nearly identical to the Efficiency,Comfort, and Satisfaction theoretical com-posites, respectively. The Absenteeism factoris made up of the absenteeism items from

Table 5Intercorrelations Among the Total Scoresof the Job Design Scales

Motivational Mechanistic Biological

MechanisticBiologicalPerceptual/

Motor

-.69*.33*

-.29*

-.06

.21* .47*

Table 6Intercorrelations Among the Theoretical JobOutcome Composites

Satisfaction Efficiency Comfort

EfficiencyComfortReliability

-.21*.27*.09

-.12*.58* .26*

Note, n = 206 * p < .05.

the satisfaction category that were not in-cluded in the Satisfaction composite. TheNonstressfulness factor is composed mostlyof a few items (e.g., work overload) from theReliability composite.

Tests of Hypotheses

The two incumbents and supervisors foreach job were randomly assigned to twosamples so that each analysis could be crossvalidated. A sample of averaged responseswas also formed. Hypotheses were tested onboth samples and the averaged sample. Coef-ficients of congruence (Wrigley & Neuhaus,1955) between the results in the two samplesand the averaged sample range from .96 to.99. Thus, only the results for the averagedsample are presented. Furthermore, each hy-pothesis was tested with the individual items,the theoretical composites, and the factorscores. Again, because the results are sohighly consistent in terms of direction andmagnitude, only those for the theoreticalcomposites are presented.5

Table 8 contains the correlations betweenthe job design scales and the theoretical com-posites. Correlations relevant to the mainhypotheses are in the diagonal from top leftto bottom right. All these correlations arepositive and significant as predicted. Regard-ing the speculative hypotheses (off-diagonal),the results in Table 8 are consistent for theMechanistic, Biological, and Perceptual/Mo-tor scales but not entirely as expected for theMotivational scale. It exhibits a very strong,

Note, n = 121 *p< .05.

4 Descriptive statistics and intercorrelations among theoutcome items are available from the first author.

5 Analyses with individual outcome items and factorscores and analyses using the cross-validation samplesare available from the first author.


Table 7Correlations Between the Theoretical Job Outcome Composites and the FactorScores of the Job Outcomes

Factor score

Note, n = 169. * p < .05.

Theoretical composite

Satisfaction Efficiency Comfort Reliability

Mental EasePhysical EaseAttitude FavorabilityAbsenteeismNonstressfulness

-.18*.14*.96*

-.03.14*

.94*-.02-.17*

.00

.06

-.03.90*.16*

-.04.21*

.66*

.24*

.07-.02

.52*

negative correlation with the Efficiency com-posite, positive correlation with the Comfortcomposite, and negative correlation with theReliability composite.

Archival measures were not included inthe theoretical composites but did correlatewith the job design scales. Occurrences absentcorrelated negatively (i.e., fewer absences)with the Motivational (ave. r via z transfor-mation = -.31) and Biological (ave. r = -.25)scales, but positively with the Mechanistic(ave. r = .23) scale. Medical incidents corre-lated negatively (i.e., fewer incidents) withthe Biological (ave. r = -.35) and Motiva-tional (ave. r = -.23) scales.

It was also predicted that each job designscale would relate to its corresponding out-come category more strongly than to anyother. Table 8 reveals some clear exceptions,especially for the Motivational and Perceptual/Motor scales and the Efficiency and Reliabilitycomposites.

Correlations between the job design scalesand the factor scores were also computed.5 Ifthe Mental Ease, Physical Ease, and AttitudeFavorability factors can be equated with theEfficiency, Comfort, and Satisfaction com-posites, respectively, the results are essentiallythe same as in Table 8. The Absenteeismfactor shows the same pattern of correlationsas did the individual items, and the Non-stressfulness factor shows no consistent cor-relations.

Because this study examinees the relation-ship between two sets of variables, a canonicalcorrelation analysis was conducted (Darling-ton, Weinberg, & Walberg, 1973; Harris,1975). This analysis is an overall multivariatetest of the hypothesis that the job design

scales are significantly related to the outcomes(Harris, 1975). Table 9 shows that two large,significant canonical correlations resulted.Two common traits or links best explain therelationship between the job design scalesand the theoretical outcome composites.

The correlations with the canonical variatesindicate which variables contribute most tothe links (Cooley & Lohnes, 1971; Darlingtonet al., 1973; Levine, 1977; Meredith, 1964).In terms of the items that make up theoutcome composites, the first canonical vari-ate shows that jobs low on the Motivationalscale but high on the Mechanistic and Per-ceptual/Motor scales tend to have high utili-zation levels, low training times, low errorlikelihoods, and low mental demands. Also,these jobs may produce less satisfaction andless physical comfort. Thus, this first linkseems to reflect the mental demands of thejob, and it is the largest of the two links.

The second canonical variate taps a phys-ical demands link. The largest correlation onthe independent variable side is for the Bio-logical scale, and the largest correlation onthe dependent variable side is for the Comfortcomposite.

Canonical analyses with the factor scoresstrongly supports the speculation that the firstcanonical variate taps a mental demandslink, whereas the second taps a physical de-mands link.7 That is, although the job designscales correlate with the variates in a nearlyidentical fashion as in Table 9, on the outcomeside the Mental Ease factor is the majorcontributor to the first variate, and the Phys-

* See Footnote 5.7 See Footnote 5.


Table 8Correlations Between the Job Design Scales and the Theoretical Job Outcome Composites

Job design scale

Note, n = 121. *p < .05.

Theoretical job outcome composite

Satisfaction Efficiency Comfort Reliability

MotivationalMechanisticBiologicalPerceptual/Motor

,32*-.22*

.15-.08

-.77*.54*

-.12.49*

.28*-.06

.50*

.01

-.49*.39*.01.45*

ical Ease factor is the main contributor tothe second variate.

A canonical analysis also indicates theamount of variance that can be explained inone set of variables given information aboutthe other set through a redundancy index(Cooley & Lohnes, 1971; Stewart & Love,1968). The redundancies in Table 9 revealthat the job design scales can account for35% of the variation in the theoretical com-posites.

Supplementary AnalysesThree types of potential moderators of the

job design-outcome relationships were ex-plored: biographical variables (e.g., sex, age,race, tenure), plant differences, and job level.The first two had no effect. Partial correlations

Table 9Results of the Canonical Correlation AnalysisBetween the Job Design Scales and theTheoretical Job Outcome Composites:Component-Variate Correlations, CanonicalCorrelations, and Redundancies

Measure

Job design scaleMotivationalMechanisticBiologicalPerceptual/Motor

Theoretical job outcomecomposite

SatisfactionEfficiencyComfortReliabilityCanonical correlationRedundancy

Canonicalvariate 1

-.94*.63*

-.25*.57*

-.36*.96*

-.37*.66*.84*.29

Canonicalvariate 2

-.07.31*.85*.31*

.04

.21*

.89*

.24*

.52*

.06

Note, n = 121. * p< .05.

while controlling for job level, denned interms of EEO code and pay, resulted in onlyslight drops in most correlations. However,controlling for job level eliminates the cor-relations between the Motivational scale andthe Satisfaction and Comfort composites, butit still remains negatively correlated with theEfficiency (e.g., partial r = -.45) and Reli-ability (e.g., partial r = -.28) composites.Furthermore, job level does not completelyexplain the correlations between the Motiva-tional scale and absenteeism (e.g., partialr=-.21).

Discussion

The four scales of the MJDQ show verygood psychometric properties, especially in-terrater reliability and agreement. Future re-search should examine the generalizability ofthe MJDQ in larger and different samples ofjobs, including nonmanufacturing and non-blue collar jobs, and further assess reliabilityby using analysts less familiar with its contentand with the jobs. If such research verifiesthe two canonical factors, then the MJDQshould be revised. Future research might alsocompare the MJDQ with other measurementinstruments, such as the Job Diagnostic Sur-vey (Hackman & Oldham, 1975) or the JobCharacteristics Inventory (Sims et al., 1975).

The intercorrelations among the scales canbe understood on rational and theoreticalgrounds. The Motivational scale has a strong,negative correlation with the Mechanistic scaledue to their diverging evaluations of featuressuch as task variety and skill usage. From atheoretical standpoint, the motivational ap-proach was a reaction against early mecha-nistic practices (L. Davis & J. Taylor, 1979).The Motivational scale has a moderate, neg-


ative correlation with the Perceptual/Motorscale. Recall that the Perceptual/Motor scalegives higher scores to jobs with fewer infor-mation processing demands, whereas the Mo-tivational scale generally scores these jobslower. The moderate, positive correlation be-tween the Motivational and Biological scalesis best understood in terms of their jointrelationship to job level.

The lack of correlation between the Mech-anistic and Biological scales is unexpectedbecause of the evidence that highly mecha-nized work may have negative health conse-quences (e.g., Johansson et al, 1978). Thislack of correlation is due to the compensatoryeffect of the muscle motion economy itemsin the Mechanistic scale. When the Mecha-nistic scale is recalculated excluding theseitems, it correlates negatively (r = —.40) withthe Biological scale. The Mechanistic andPerceptual/Motor scales correlate positivelybecause they both highly score jobs withfewer mental demands. This also makes sense,as industrial engineering is a major contrib-utor to the field of human factors (Meister,1971; Meister & Rabideau, 1965).

The Biological and Perceptual/Motor scalesalso positively correlate probably because oftheir joint concern for proper person-machinefit. Although the former is concerned withbiological fit and the latter with perceptual/motor fit, the modern-day practice of humanfactors or ergonomics includes both consid-erations (e.g., Grandjean, 1980).

To simplify the presentation of the results,the pool of outcomes was reduced to two setsof scales: four theoretical composites and fivefactor scores. Development and analyses ofthese scales yields three types of evidencesupporting the accuracy of the outcome tax-onomy. First, it is possible to form an inter-nally consistent composite within each cate-gory that includes most of the hems. Second,the empirical clustering of the outcome itemsvia factor analysis largely reproduces three ofthe four theoretical composites. Third, thepattern of intercorrelations among the theo-retical composites is similar to the pattern ofintercorrelations among the correspondingjob design scales. For example, comparingTables 6 with 5 shows that the Satisfactioncomposite correlates negatively with the Ef-ficiency and positively with the Comfort com-

posites. Similarly, the Motivational designscale correlates negatively with the Mecha-nistic and positively with the Biological scales.These results suggest an empirical as well asa theoretical symmetry between the job designand outcome taxonomies.

When the hypotheses were tested, the re-sults were consistent across the various tech-niques, samples, and outcome combinations.In general, the main hypotheses are wellsupported. Jobs that score high on the Mo-tivational scale have employees who are moresatisfied and motivated, have higher rated jobperformance, and exhibit less absenteeism.Jobs high on the Mechanistic scale havehigher utilization levels and lower trainingrequirements. Jobs high on the Biologicalscale require less physical effort, producefewer aches and pains, and result in fewermedical incidents. Finally, jobs high on thePerceptual/Motor scale are less likely to pro-duce accidents, errors, stress, and work over-load, and require fewer mental demands.

With few exceptions, most of the specula-tive hypotheses are also supported. Jobs withmore motivational features require moretraining time and have lower utilization levels.Contrary to predictions, jobs with more mo-tivational features have lower effort require-ments, greater comfort, and fewer healthcomplaints. But this finding may be the spu-rious result of their joint relationship withjob level. Also unexpected, high motivationaljobs have greater accident and error likeli-hoods, more stress and overload, and moremental demands. Although these variablesalso correlate with job level, more motivatingjobs may simply have more mental demandsand greater chances of error and overload ingeneral.

No other major exceptions to the specula-tive hypotheses occurred. Jobs high on theMechanistic scale may have less satisfied em-ployees and slightly higher absenteeism, butthey tend to be less accident and error likelyand less prone to mental overload. Mechanis-tic design shows no relationship with physicaloutcomes. But when the Mechanistic scale isrecalculated excluding the muscle motioneconomy items, it correlates negatively withphysical outcomes (e.g., ave. r = —.28). Thebiological features of a job are unrelated toany of the outcomes, aside from the physical


outcomes. Finally, perceptual/motor charac-teristics are unrelated to attitudes, effort, orhealth but are positively associated with higherutilization levels and less training time.

It was also predicted that each job designscale would relate most strongly to its corre-sponding outcome category. This predictionwas not supported. Recall, however, that thejob design scales were developed based on acontent evaluation of current theories andnot on an empirical clustering scheme. Ifadditional research on larger and more diversesamples of jobs yields comparable results,these unexpected findings should lead toreexamination of theories rather than to re-naming of scales.

Canonical correlation analyses showed thattwo links explain most of the relationshipamong the job design scales and outcomes.The largest link reflects the mental demandsof the job, whereas the smaller reflects aphysical demands component. This suggestsa simplified two-factor schema of the influ-ence of job design on important outcomes.The large portion of the variation that thejob design scales can account for in theoutcomes indicates the practical as well asthe theoretical importance of the results.

When potential moderators of the job de-sign-outcome relationships were explored, joblevel reduced the magnitudes of most of thecorrelations slightly, especially those betweenthe Motivational scale and the Satisfactionand Comfort composites. But the fact thathigher level jobs are typically more satisfyingand motivating and less physically demandingdoes not diminish the theoretical importanceof the scale or the previous findings for manyreasons. First, it is more likely that a job'scharacteristics determine the job's level andpay, rather than the reverse. Second, job leveland pay would be poor substitutes for theMotivational scale, as they yield little infor-mation as to why jobs are satisfying andmotivating. Finally, even when job level andpay are partialed out, the scale still correlateswith absenteeism.

Theoretical Implications

This study demonstrates that different ap-proaches to job design can be reliably mea-sured in a field setting, and they relate to

important outcomes for both the individualand the organization. As no single approachcan fully explain all outcomes, an interdisci-plinary perspective is suggested. The job de-sign taxomony derived may have merit forthis purpose. It reflects the content of allcurrent theories, and it has criterion-relatedvalidity with the outcomes. The MJDQ op-erationally defines the taxonomy for purposesof research and application.

The outcome taxonomy may also be auseful way to conceptualize outcomes fromjob design. It reflects the theoretical frame-work of the design taxonomy and showssimilar internal relationships. It is empiricallysupported by a factor analysis of a large setof outcomes. Finally, much of the variationin the outcomes is explained by the scales ofthe MJDQ.

Although there is some overlap betweenthe job design approaches, there are alsosome basic conflicts. Most of these differencesare between the motivational approach andthe perceptual/motor and mechanistic ap-proaches. Although the motivational and per-ceptual/motor approaches were both derivedfrom psychology, they intercorrelate negativelyand show some negative relationships witheach other's outcomes. Clearly, each approachhas a different orientation. The perceptual/motor approach strives to develop equipmentand jobs that are simple, safe, reliable, andminimize the mental demands required ofworkers. Conversely, the motivational ap-proach stresses that more complicated andchallenging jobs are more rewarding andshould be encouraged.

The most glaring conflict is between themotivational and mechanistic approaches.They make nearly opposite recommendationsin terms of job complexity and mental de-mands, and they show many negative corre-lations with each other's outcomes. It is cu-rious that so many authors write of thenegative consequences for individuals ofmechanistic designed jobs, but few commenton the costs of the motivational approach interms of important organizational outcomessuch as utilization levels, training times, ac-cident potential, and error likelihood.

These conflicts may be partially resolved,however. A job could gain on one approachwithout sacrificing its status on others, but


trade offs will probably be necessary. As totrade offs, mental and physical demands ofjobs seem to be relatively independent. Phys-ical demands of jobs can probably be reducedwithout sacrificing the jobs' mental demands.The major trade offs will most likely involvemental demands. One might conceive of amental demands continuum with motivation-ally designed (mentally demanding) jobs onone end that maximize individual outcomeslike high satisfaction and high motivation.On the other end are mechanistic and per-ceptual/motor designed (less mentally de-manding) jobs that maximize such organiza-tional outcomes as high utilization levels,short training times, and low error likelihoods.Which trade offs will be made depend onwhich outcomes one wants to maximize, andthe choice of outcomes depends on one'svalues (L. Davis, Canter, & Hoffman, 1955;J. Taylor, 1979). Trade offs will depend partlyupon how one values individual versus orga-nizational outcomes.

In summary, findings of this study mayserve as a vehicle for rethinking major theoriesof job design. The taxonomies, measures, andfindings of this study may help clarify thesimilarities and differences among the ap-proaches, delineate the costs and benefits ofeach approach, rectify or find compromisesfor apparent conflicts among some of theapproaches, and integrate the approaches intoa comprehensive interdisciplinary theory ofjob design.

Practical Implications

Most of the practical implications of thisstudy relate to the use of the MJDQ inapplied job design research. The MJDQ isan easy-to-use, analytical aid for the researcher(e.g., psychologist, engineer, ergonomist, man-ager, technician). It might encourage the ex-amination of job design as an importantorganizational variable and provide an inte-grated and structured means of doing so.

At least three types of practical uses couldbe made of the MJDQ. First, when organi-zational problems occur, it could be used todetermine if problems exist with the designof the jobs. For example, the authors encoun-tered a situation where an employee wasbeing reprimanded for poor performance,

whereas the MJDQ revealed that the jobcreated such extreme biomechanical stressesin the back and legs such that it was nearlyimpossible to perform satisfactorily for anyextended period. Second, the MJDQ may beuseful in job redesign projects to help identifyjobs that need redesign, to tell the nature ofthe redesign needed, and to evaluate jobsafter they have been modified. Third, theMJDQ would be useful during many of thephases of system development: as a guide forjob design recommendations in the designphase; as a checklist for evaluating equipment,workplace, and other design prototypes duringthe development phase; and as an evaluationinstrument once the system is fully developed.In short, the MJDQ would be a useful aidfor the designer to help assure that bothmental and physical needs and limitations ofpeople are recognized. It may be possible todesign jobs that are optimal from all perspec-tives, thus satisfying the needs of both theindividual and the organization.

ReferencesArgyris, C. (1964). Integrating the individual and the

organization. New York: Wiley.Astrand, P. O., & Rodahl, K. (1977). Textbook of work

physiology: Physiological bases of exercise (2nd ed.).New York: McGraw-Hill.

Ayoub, M. M. (1973). Work place design and posture.Human Factors, 15, 265-268.

Bare, C. E. (1966). The measurement of attitudes towardman-machine systems. Human Factors, 8, 71-79.

Barnes, R. M. (1980). Motion and time study: Designand measurement of work (7th ed.). New York: Wiley.

Bartley, S. H., & Chute, E. (1947). Fatigue and impairmentin man. New York: McGraw-Hill.

Bass, B. M., Cascio, W. E, & O'Connor, E. J. (1974).Magnitude estimations of expressions of frequency andamount. Journal of Applied Psychology, 59, 313-320.

Borg, G. (1962). Physical performance and perceivedexertion. Lund, Sweden: Gleerup.

Brayfield, A. H., & Rothe, H. E (1951). An index of jobsatisfaction. Journal of Applied Psychology, 35, 307-311. ,

Caldwell, D. F, & O'Reilly, C. A. (1982). Task perceptionsand job satisfaction: A question of causality. Journalof Applied Psychology, 67, 361-369.

Campbell, D. T, & Fiske, D. W. (1959). Convergent anddiscriminant validation by the multitrait-multimethodmatrix. Psychological Bulletin, 56, 81-105.

Caplan, R. D., Cobb, S., French, J. R. P., Van Harrison,R., & Pinneau, S. R. (1975). Job demands and workerhealth: Main effects and occupational differences (HEWPublication No. [NIOSH] 75-160). Washington, DC:U.S. Government Printing Office.

Cascio, W. E (1978). Applied psychology in personnelmanagement. Reston, VA: Reston.


Chaffin, D. B. (1974). Human strength capability andlow-back pain. Journal of Occupational Medicine, 16,248-254.

Cherns, A. (1976). The principles of sociotechnical design.Human Relations, 29, 783-792.

Cohen, J. (1977). Statistical power analysis for the be-havioral sciences (Rev. ed.). New York: AcademicPress.

Cooley, W. W., & Lohnes, P. R. (1971). Multivariatedata analysis. New York: Wiley.

Corlett, E. N., & Bishop, R. P. (1976). A technique forassessing postural discomfort. Ergonomics, 19, 175-182.

Cronbach, L. J. (1951). Coefficient alpha and the internalstructure of tests. Psychometrika, 16, 297-334.

Darlington, R. B., Weinberg, S. L., & Walberg, H. J.(1973). Canonical variate analysis and related tech-niques. Review of Educational Research, 43, 433-454.

Davis, H. L., & Miller, C. I. (1971). Human productivityand work design. In H. B. Maynard (Ed.), Industrialengineering handbook (3rd ed., pp. 7.74-7.101). NewYork: McGraw-Hill.

Davis, L. E., Canter, R. R., & Hoffman, J. (1955).Current job design criteria. Journal of Industrial En-gineering, 6(2), 5-8; 21-23.

Davis, L. E., & Taylor, J. C. (Eds.). (1979). Design ofjobs (2nd ed.). Santa Monica, CA: Goodyear.

Dunnette, M. D. (Ed.). (1976). Handbook of industrialand organizational psychology. Chicago: Rand-McNally.

Einhorn, H. J., & Hogarth, R. M. (1975). Unit weightingschemes for decision-making. Organizational Behaviorand Human Performance, 13, 171-192.

Englestad, P. H. (1979). Sociotechnical approach toproblems of process control. In L. E. Davis & J. C.Taylor (Eds.), Design of jobs (2nd ed., pp. 184-205).Santa Monica, CA: Goodyear.

Floyd, W. F., & Ward, J. (1966). Posture in industry.International Journal of Production Research, 5, 213-224.

Fogel, L. J. (1967). Human information processing. En-glewood Cliffs, NJ: Prentice-Hall.

Ford, R. N. (1969). Motivation through the work itself.New York: American Management Association.

Frankenhaeuser, M. (1977). Job demands, health andwell-being. Journal of Psychosomatic Research, 21,313-321.

Gagne, R. M. (1962). Human functions in systems. InR. M. Gagne (Ed.), Psychological principles in systemdevelopment (pp. 35-73). New York: Holt, Rinehart,and Winston.

Gilbreth, F. B. (1911). Motion study: A method forincreasing the efficiency of the workman. New York:Van Nostrand.

Goodstein, L. P. (1981). Discriminative display supportfor process operators. In J. Rasmussen & W. B. Rouse(Eds.), Human detection and diagnosis of system failures(pp. 241-258). New York: Plenum.

Grandjean, E. (1980). Fitting the task to the man: Anergonomic approach (3rd ed.). London: Taylor &Francis.

Grandjean, E., & Hunting, W. (1977). Ergonomics ofposture—Review of various problems of standing andsitting posture. Applied Ergonomics, 8, 135-140.

Griffin, R. W. (1982). Task design: An integrative ap-proach. Glenview, IL: Scott-Foresman.

Hackman, J. R., & Lawler, E. E. (1971). Employeereactions to job characteristics. Journal of AppliedPsychology, 55, 259-286. (Monograph).

Hackman, J. R., & Oldham, G. R. (1975). Developmentof the job diagnostic survey. Journal of Applied Psy-chology, 60, 159-170.

Hackman, J. R., & Oldham, G. R. (1980). Work redesign.Reading, MA: Addison-Wesley.

Hammer, T. H., & Landau, J. (1981). Methodologicalissues in the use of absence data. Journal of AppliedPsychology, 66, 574-581.

Harris, R. J. (1975). A primer of multivariate statistics.New York: Academic Press.

Hertzberg, H. T. H. (1972). Engineering anthropology.In H. P. Van Cott & R. G. Kinkade (Eds.), Humanengineering guide to equipment design (Rev. ed., pp.467-584). Washington, DC: U.S. Government PrintingOffice.

Herzberg, F. (1966). Work and the nature of man. Cleve-land: World.

Herzberg, F. (1968). One more time: How do you motivateemployees? Harvard Business Review, 48(1), 53-62.

Hulin, C. L., & Blood, M. R. (1968). Job enlargement,individual differences, and worker responses. Psycho-logical Bulletin, 69, 41-55.

Ireson, W. G., & Grant, E. L. (Eds.). (1971). Handbookof industrial engineering and management (2nd ed.).Englewood Cliffs, NJ: Prentice-Hall.

Ivancevich, J. M., & Matteson, M. T. (1980). Stress andwork: A managerial perspective. Glenview, IL: Scott-Foresman.

Jenkins, G. D., Nadler, D. A., Lawler, E. E., & Cammann,C. (1975). Standardized observations: An approach tomeasuring the nature of jobs. Journal of AppliedPsychology, 60, 171-181.

Johansson, G., Aronsson, G., & Lindstrom, B. O. (1978).Social psychological and neuroendocrine stress reactionsin highly mechanized work. Ergonomics, 21, 583-599.

Kahneman, D. (1973). Attention and effort. EnglewoodCliffs, NJ: Prentice-Hall.

Karasek, R. A. (1979). Job demands, job decision latitude,and mental strain: Implications for job redesign. Ad-ministrative Science Quarterly, 24, 285-308.

Kinsman, R. A., & Weiser, P. C. (1976). Subjectivesymptomatology during work and fatigue. In E. Si-monson & P. C. Weiser (Eds.), Psychological aspectsand physiological correlates of work and fatigue (pp.336-405). Springfield, IL: Charles C. Thomas.

Konz, S. (1979). Work design. Columbus, OH: Grid.Kornhauser, A. (1965). Mental health of the industrial

worker: A Detroit study. New York: Wiley.Levine, M. S. (1977). Canonical analysis and factor

comparison. Sage University paper series on quantitativeapplications in the social sciences, 07-006. BeverlyHills: Sage.

Likert, R. (1961). New patterns of management. NewYork: McGraw-Hill.

Lodahl, T. M., & Kejner, M. (1965). The definition andmeasurement of job involvement. Journal of AppliedPsychology, 49, 24-33.

Mayer, J. R. (1971). New perspectives in job enrichment.New York: Van Nostrand Reinhold.


Maynard, H. B. (Ed.). (1971). Industrial engineeringhandbook (3rd ed.). New York: McGraw-Hill.

McCormick, E. J. (1976). Human factors in engineeringand design (4th ed.). New \brk: McGraw-Hill.

McGregor, D. (1960). The human side of enterprise. NewYork: McGraw-Hill.

Meister, D. (1971). Human factors: Theory and practice.New York: Wiley.

Meister, D., & Rabideau, G. F. (1965). Human factorsevaluation in system development. New York: Wiley.

Meredith, W. (1964). Canonical correlations with fallibledata. Psychometrika, 29, 55-65.

Mitchell, T. R. (1976). Applied principles in motivationtheory. In P. Warr (Ed.), Personal goals in work design(pp. 163-171). New York: Wiley.

Morgan, C. T., Cook, J. S., Chapanis, A., & Lund,M. W. (Eds.). (1963). Human engineering guide toequipment design. New \brk: McGraw-Hill.

Mundel, M. E. (1970). Motion and time study: Principlesand practices (4th ed.). Englewood Cliffs, NJ: Prentice-Hall.

Nadler, G. (1963). Work design. Homewood, IL: Irwin.O'Reilly, C. A., & Caldwell, D. F. (1979). Informational

influence as a determinant of perceived task charac-teristics and job satisfaction. Journal of Applied Psy-chology, 64, 157-165.

O'Reilly, C. A., Parlette, G. N., & Bloom, J. R. (1980).Perceptual measures of task characteristics: The biasingeffects of differing frames of reference and job attitudes.Academy of Management Journal, 23, 118-131.

Park, K. S., & Chaffin, D. B. (1975). Prediction of load-lifting limits for manual materials handling. ProfessionalSafety, 20(5), 44-48.

Pearson, R. G. (1957). Scale analysis of a fatigue checklist.Journal of Applied Psychology, 41, 186-191.

Pierce, J. L., & Dunham, R. B. (1976). Task design: Aliterature review. Academy of Management Review,1(4), 83-97.

Rasmussen, J. (1981). Models of mental strategies inprocess plant diagnosis. In J. Rasmussen & W. B.Rouse (Eds.), Human detection and diagnosis of systemfailures (pp. 433-449). New York: Plenum.

Roberts, K. H., & Click, W. (1981). The job characteristicsapproach to task design: A critical review. Journal ofApplied Psychology, 66, 193-217.

Roebuck, J. A., Kroemer, K. H. E., & Thompson, W. G.(1975). Engineering anthropometry methods. New \brk:Wiley.

Rousseau, D. M. (1977). Technological differences in jobcharacteristics, employee satisfaction, and motivation:A synthesis of job design research and sociotechnicalsystems theory. Organizational Behavior and HumanPerformance, 19, 18-42.

Salvendy, G., & Smith, M. J. (Eds.). (1981). Machinepacing and occupational stress. London: Taylor &Francis.

Shaw, J. B. (1980). An information-processing approachto the study of job design. Academy of ManagementReview, 5, 41-48.

Shepard, J. M. (1969). Functional specialization andwork attitudes. Industrial Relations, 8, 185-194.

Shepard, J. M. (1970). Functional specialization, alien-ation, and job satisfaction. Industrial and Labor Re-lations Review, 23, 207-219.

Sims, H. P., Szilagyi, A. D., & Keller, R. T. (1976). Themeasurement of job characteristics. Academy of Man-agement Journal, 19, 195-212.

Snook, S. H., & Irvine, C. H. (1967). Maximum accept-able weight of lift. American Industrial Hygiene As-sociation Journal, 28, 322-329.

Steers, R. M., & Mowday, R. T. (1977). The motivationalproperties of tasks. Academy of Management Review,2, 645-658.

Stewart, D., & Love, W. (1968). A general canonicalcorrelation index. Psychological Bulletin, 70, 160-163.

Swain, A. (1973). An error-cause removal program forindustry. Human Factors, 15, 207-221.

Taylor, F. W. (1911). The principles of scientific manage-ment. New %rk: Norton.

Taylor, J. C. (1979). Job design criteria twenty years later.In L. E. Davis & J. C. Taylor (Eds.), Design of jobs(2nd ed., pp. 54-63). Santa Monica, CA: Goodyear.

Tichauer, E. R. (1978). The biomechanical basis ofergonomics: Anatomy applied to the design of worksituations. New \brk: Wiley.

Turner, A. N., & Lawrence, P. R. (1965). Industrial jobsand the worker: An investigation of response to taskattributes. Boston: Harvard Graduate School of Busi-ness Administration.

Van Cott, H. P., & Kinkade, R. G. (Eds.). (1972). Humanengineering guide to equipment design (Rev. ed.).Washington, DC: U.S. Government Printing Office.

Van Wely, P. (1970). Design and disease. Applied Ergo-nomics, 1, 262-269.

Vroom, V. H. (1964). Work and motivation. New York:Wiley.

Wainer, H. (1976). Estimating coefficients in linear models:It don't make no nevermind. Psychological Bulletin,83, 213-217.

Walker, C. R., & Guest, R. H. (1952). The man on theassembly line. Cambridge: Harvard University Press.

Walters, R. W. (1975). Job enrichment for results: Strat-egies for successful implementation. Reading, MA:Addison-Wesley.

Weber, A., Fussier, C., O'Hanlon, J. E, Gierer, R,, &Grandjean, E. (1980). Psychophysiological effects ofrepetitive tasks. Ergonomics, 23, 1033-1046.

Welford, A. T. (1976). Skilled performance: Perceptualand motor skills. Glenview, IL: Scott-Foresman.

Wexley, K. N., & Yukl, G. A. (1977). Organizationalbehavior and personnel psychology. Homewood, IL:Irwin.

Woodson, W. E. (1954). Human engineering guide forequipment designers. Berkeley, CA: University of Cal-ifornia Press.

Woodson, W. E. (1981). Human factors design handbook.New York: McGraw-Hill.

Woodson, W. E., & Conover, D. W. (1964). Humanengineering guide for equipment designers (2nd ed.).Berkeley, CA: University of California Press.

Wrigley, C., & Neuhaus, J. (1955). The matching of twosets of factors. American Psychologist, 10, 418-419.


Appendix

Sample Items From the Multimethod Job Design Questionnaire (MJDQ)

Sample Item From the Motivational JobDesign Scale

(#1) Autonomy, responsibility, vertical loading:Does the job allow autonomy? To what extentdoes the job allow freedom, independence, ordiscretion in work scheduling or sequence, workmethods or procedures, or quality control, etc.?How much control or responsibility for decisionmaking concerning the work does this job allow?Is the job vertically loaded?5. The job allows almost complete autonomy in

work sequencing, methods, etc. Employee hasalmost complete responsibility for decisionmaking concerning the work.

3. The job allows some autonomy and responsi-bility for decision making, but some of thedecisions are fixed or made by supervisors orothers within the organization.

1. The job allows very little autonomy and respon-sibility for decision making. Almost all decisionsconcerning scheduling, methods, procedures,etc., are fixed or made by others.

(Note: The following relevance scale and commentsline are included for each item of the MJDQ.)3. Highly relevant2. Minimally relevant1. Irrelevant (not rated)Comments:

Sample Item From the Mechanistic JobDesign Scale

(#4) Skill simplification: To what extent is thejob designed in such a way that it requires as littleskill and training time as possible? To what extentcan nearly anyone perform the job with littlepractice?5. The job requires very little skill and training

time. Most anyone can perform the job withlittle practice. Training time is only from a fewhours to a few days.

3. The job requires only moderate amounts ofskill and training time. Complete mastery ofthe job takes from a few weeks to a few months.

I. The job requires a great deal of skill andtraining time. Training time for this job/skilltakes from a year to a few years.

Sample Item From the Biological JobDesign Scale

(#6) Strength: To what extent are the muscularstrength requirements of the job reasonable? Do

any of the tasks require strength levels that mayexceed the capabilities of the workers required toperform them? Aside from maximum strengthlevels required, also consider the continuous versusintermittent nature of the tasks.5. This job requires only a limited amount of

muscular strength. Only minimal strength isrequired for continuous tasks, and only mod-erate strength is required for intermittent tasks;strength levels required would not exceed theaverage capability level of the general population.

3. The job requires a moderate amount of mus-cular strength. Only moderate strength is re-quired for continuous tasks, and higher levelsof strength are required only intermittently;strength levels required would slightly exceedthe average capability level of the general pop-ulation.

1. The job requires a great amount of muscularstrength. High levels of strength are requiredon a continuous basis, and/or excessive levelsof strength are required intermittently; strengthlevels required would greatly exceed the averagecapability level of the general population.

Sample Item From the Perceptual/Motor JobDesign Scale

(#15) Input requirements: To what extent arethe information input requirements on the jobwithin the limitations of the least capable potentialworker? Considering all forms of information thatmust be sensed and perceived to effectively performthe job, does the quantity of information, the ratepresented, the quality (e.g., discriminability) of thestimuli, etc., result in job requirements that couldbe met by the least capable potential worker?5. The information input requirements on this

job are minimal. They are within the capabilitiesof nearly all potential workers and require littlemental effort or training/experience.

3. The information input requirements on thisjob are moderate. They are within the capabil-ities of the average potential worker but requiresome mental effort and/or training/experience.

1. The information input requirements on thisjob are considerable. They are within the ca-pabilities of only the above average potentialworker and require much mental effort and/ortraining/experience.

Received November 22, 1983Revision received May 15, 1984 •

Development and Field Evaluation of an Interdisciplinary ... · job enrichment and motivating job...

Documents

Transcript of Development and Field Evaluation of an Interdisciplinary ... · job enrichment and motivating job...