A RE-INQUIRY OF HOFSTEDE'S CULTURAL DIMENSIONS: A CALL …

A Re-Inquiry of Hofstede's Cultural Orr and Häuser

A RE-INQUIRY OF HOFSTEDE'S CULTURAL DIMENSIONS:A CALL FOR 21st CENTURY

CROSS-CULTURAL RESEARCHLINDA M. ORR, The University of Akron

WILLIAM J. HÄUSER, The University of Akron

Given the impact of Hofstede 's Cultural Dimensions over the past quarter of a century, manyscholars and practitioners have utilized these dimensions. However, numerous researchers havequestioned his methodology, while others misused the dimensions in terms of the original purpose.Yet surprisingly, very few studies have performed an exact replication. This study summarizesHofstede's work and critiques his cross-cultural model. In order to test Hofstede's constructs ondifferent populations, three quantitative analyses were performed using domestic U.S., Asian, andAustralian samples. This study found serious problems with Hofstede's factor structure. Additionally,the study suggests the need for re-examining the cultural dimensions within the global informationbased context of the early 21^' century. This is not meant to criticize Hofstede, but instead to pinpointfallacies to enable researchers to build from his work in more appropriate directions.

INTRODUCTION

In 1980, Geert Hofstede published Culture'sConsequences. This influential study soonbecame a major source of reference about valuedifferences around the world. Culture'sConsequences has been translated intonumerous languages since its originalpublication and was fully revised in 2001.Additionally, Culture's Consequences has beencited more than any other book in socialsciences (Yoo and Dunthu 2002). Hofstede hasbeen cited over 5,000 times, with more than3,000 of these citations being to Culture'sConsequences (IACCP 2007). Hofstede's workhas made it beyond higher-level research andhas worked its way into everyday teachings.Many of his citations are in basic principles ofmarketing, international business, advertising,and consumer behavior textbooks.

Hofstede's work has inspired a multitude ofinternational marketing research and has beenthe dominant research paradigm in cross-cultural studies of national attitudes for sometime because of the simplicity of its theory and

The Marketing Management JournalVolume 18, Issue 2, Pages 1-19Copyright O 2008, The Maiketing Management AssoeiationAll rights of rcproducdon in any form reserved

applicability of its implications. Hofstede'swork has inspired a great improvement in thediscipline by specifying a theoretical modelwhich serves to coordinate research efforts(Redding 1995). This theoretical model hasserved as the foundation for many otherresearch efforts. In sum, Hofstede's initial four(later five) fundamental dimensions of culturestill serve today as basic, fundamental criteriain most interdisciplinary, cross-culturallycomparative research.

However, Hofstode's work has beenmisconstrued and misinterpreted in manysubsequent studies. More importantly,surprisingly few exact replications, attemptingto empirically examine Hofstede's factors, havebeen conducted. Furthermore, many subsequentstudies have taken and utilized Hofstede's workwith surprisingly little questioning of hisresults. Thus, due to the tremendous impact thatHofstede has had on the scientific community,an exact replication study is a necessary step.While replication studies may not hold thehighest regard within academic researchbecause they do not bring "anything new to thetable," replication studies for works thisimportant must be done. If not, as researchers,we will continue to utilize faulty theories andmodels. Thus, a re])lication study of Hofstede's

Marketing Management Journal, Fall 2008


work is necessary in order to assess theappropriate content, as well as the reliabilityand validity of Hofstede's cultural dimensions.

Thus, the purpose of this study is to perform avery close replication of Hofstede's originalstudy. To do this, first, a review of Hofstede'smethodology is presented. This review wasconducted because many researchers may notbe fully aware of the almost haphazard mannerin which the cultural dimensions were initiallydeveloped. The results of three studies designedto analyze face validity are preformed.Following these analyses, three re-inquirystudies are presented. Hofstede's instrumentwas obtained and reproduced with noalterations and administered with multipleinternational samples. The examination of thedata in this study was conducted with morestatistical rigor than any other replication studyknown to the authors. As we have progressedwell into the new millennium, cross-culturalresearch is no doubt one of the most importantsubsections of marketing and internationalbusiness research. It is imperative that we havea solid foundation and understanding fromwhich to build future research.

BACKGROUND

As mentioned, Hofstde's famous study iswidely recognized as a major break-through incross-cultural social science studies. There arealmost no publications, either from thedisciplines of sociology, anthropology, history,law, economics or business administration, thatdo not refer to Hofstede's work and his fivefundamental dimensions of culture whenexplaining correspondences and distinctionsbetween cultures (IRIC online 2002). GeertHofstede is among the 20 most cited Europeansin the 2000 Social Science Citation Index(Institute for Research on InterculturalCooperation 2001), 57* in the world, with 416articles referring to him. In fact, Hofstede'sinfluence is becoming even more pronounced,with the number of citations increasing, notdecreasing, each subsequent year. Simply put,Hofstede's dimensions are still utilized widelyeven as we have progressed well into the new


millennium. Thus, given the widespreadacceptance of Hofstede's instrument, it wasused for this research.

Hofstede's work is based on "mentalprograms." Due to the process of socialization,these mental programs are developed in thefamily in early childhood and reinforced inschools and organizations, and other areasthroughout our lives, experiences, andupbringings. Thus, due to the shared commonexperiences of people living in the samecountry, these mental programs contain acomponent of national culture. They are mostclearly expressed in the different values thatpredominate among people from differentcountries (Hofstede 1980).

From 1967 to 1972, Hofstede administered117,000 questionnaires to employees of IBM inover 60 different countries (Hofstede 1980).His study resulted from the collaboration ofresearchers from five countries and his surveywas pre-tested in ten countries (De Cieri andDowling 1995). By 1980, he had developed hisown cultural dimensions, Individualism-Collectivism, Power Distance, UncertaintyAvoidance, and Masculinity-Femininity.

Power distance is defined as the degree thatunequal distributions of power are expected andaccepted. Power distance "represents a nation'sunique score on how to deal with socialinequality. Inequality can occur in areas such asprestige, wealth, and power; different societiesput different weights on status consistencyamong these areas" (Hofstede 1984, p.65). Uncertainty avoidance is the extent towhich people feel threatened by ambiguoussituations and have created beliefs andinstitutions that try to avoid these (Hofstede andBond 1984, p. 419-420). Individualism-Collectivism "describes the relationshipbetween the individual and the collectivitywhich prevails in a given society," where"individualism is defined as a situation inwhich people are supposed to look afterthemselves" and "collectivism is defined as asituation in which people belong to in-groups orcollectivities which are supposed to look after


them in exchange for loyalty" (Hofstede 1984,p. 148 and Hofstede and Bond, 1984, p. 419-420). Finally, masculinity-femininity "describesthe division of social roles between women andmen in a society." The predominantsocialization pattern is for men to be moreassertive and for women to be more nurturing."Masculinity is defined as situation in whichthe dominant values in society are success,money, and things" and "femininity is definedas a situation in which the dominant values insociety are caring for others and the quality oflife" (Hofstede 1984, p. 176; Hofstede andBond 1984, p. 419-420).

A group of researchers calling themselves theChinese Culture Connection (1987) conductedfurther analysis of the Hofstede dimensions inAsian cultures and added a fifth dimension,Confucian Dynamism. In the 2001 edition ofCulture's Consequences, Hofstede has includeda chapter on this dimension and called it long-term versus short-term orientation. ConfucianDynamism conceptually incorporates manydiverse elements of Confucian cultures.Empirically, however, Confucian Dynamismhas consisted of two negatively correlated setsof items, described by the Chinese CultureConnection as a positive and a negative pole.More specifically, "there were four positivelyloaded values in this grouping, all reflectingConfucian work ethics." These four items were,ordering relationships, thrift, persistence, and asense of shame - all represented by singleitems. "Counterpointed against this hierarchicaldynamism were four negatively loaded valuesadvocating checks and distraction at thepersonal, interpersonal, and social levels"(Chinese Cultural Connection 1987, p. 150).These items were, reciprocation, personalsteadiness, protecting your face, and respect fortradition. Thus, Confucian Dynamism begins toaddress traditional eastern values.

HOFSTEDE'S METHODOLOGY

IBM had occasionally surveyed employees tojudge attitudes toward job satisfaction prior to1960. In 1967, a team of researchers wasgathered together to standardize the surveys in

order to permit longitudinal and cross-nationalinvestigation. The first instrument consisted of180 items, which were chosen through existingsurveys, pilot studies, and literature review(e.g., Baehr 1954; Herzberg et al. 1957;Hinrichs 1968; Vroom 1964; Wherry 1954).After the initial sui vey, individual site surveyadministrators were still customizing thesurveys to their spixific needs. Thus, most ofthe surveys at this point varied considerablyfrom site to site. Therefore, a 1970 task force ofresearchers, including Hofstede, took over witha new approach. They wanted to derive aninstrument that usetl the previous questions buthad no more than 60 items. The criteria forthese questions were:The core questions should cover all the majorarea or dimensions of job attitudes (contentvalidity);• The areas of job attitudes covered should be

meaningful in terms of theories of humanmotivation and organization;

• The questions should be reasonablyreliable,

• Core questionnaire items should universallyapply to all employees of the corporation;

• Questions should be translatable;• The questions should be chosen from those

used before, to permit longitudinal studies;• Questions should be acceptable to the

corporation's managers;• All core questions should be useful

information to managers; and• The number should not exceed 60 items

(Hofstede 1975).

In 1971, Hofstede and colleagues reduced thenumber of items from 180 to 120. His decisionwas to eliminate il ems that did not frequentlyappear in the litenture. Hofstede next ran anexploratory factor analysis with a sample of700 employees on ull 120 items. After varimaxrotation, he was left with 15 factors. The firstthree factors explained 77 percent of thevariance. Therefore, he kept the items for thosethree factors and eliminated the other items. Hewas left with three dimensions: management,satisfaction, and culture. Hofstede andcolleagues (1971) had 146 items at this point,including demogra])hic variables. Next, he took



his new survey with 146 items andadministered it to 5 separate populations, whichfollow.• Technical experts, France, 1968, n = 436• Technical experts, U.K., 1968, n = 436• Head office clerks, secretaries and other

nonprofessional employees, U.K., 1969, n= 535

• Unskilled direct manufacturing operators,Japan, 1970, n = 231

• Unskilled direct manufacturing operators,U.K., 1970, n = 296

To analyze the data, Hofstede preformedseparate factor analyses with each population.Management questions explained 21-27 percent(depending on the sample) of the variance andthe weakest factor, culture, explained 11-19percent of the variance (see Tables 1 and 2).However, the three components that Hofstedederived, management, satisfaction, and culture,were not one-dimensional. In fact, the "culture"components contained thirteen constructs, someof which contained zero or only one item. Inother words, even though there was no data or

quantitative evidence to refer to, Hofstedeconceptually decided which components shouldbe a part of each of the three factors.

From the previous analyses, Hofstede devisedhis new instrument. The final questionnaireencompassed 60 core questions: 58 from factoranalysis and two new items. Subsequently,Hofstede administered the same worksatisfaction survey in other countries andderived his "cultural" dimensions. Thefollowing paragraphs will describe how each ofthe Hofstede cultural dimensions, as we knowthem today, are derived through the use oftheoretical reasoning and factor analysis.

Hofstede began administering the instrument inindividual countries at this stage. Each factoranalysis was preformed separately for eachcountry and then standardized, normalizedmeans were calculated to derive a factor score.Hofstede himself admits that factor structuredoes not hold across populations (Hofstede1984, p. 43). In fact, as mentioned, he neverintended his instrument to be used at the

TABLE 1Hofstede et al. (1971) Factor Analysis

Satisfaction questions:1.2.3.4.5.

T. E. FranceT. E. U.K.Clerks U.K.Operators JapanOperators U.K.

Management questions:1.2.3.4.5.

T. E. FranceT. E. U.K.Clerks U.K.Operators JapanOperators U.K.

Culture questions:1.2.3.4.5.

T. E. FranceT. E. U.K.Clerks U.K.Operators JapanCurators U.K.

Number ofvariables

5454525154

5050504850

4242404042

Number of factorswith eigen-values>l

1515151515

1414131514

1214111311

% of total vari-ance explained

6262636865

6261637265

5861596359

% of varianceexplained byfirst factor

1716202420

2221242724

1311171619



TABLE 2Hofstede's Factors

SatisfactionFactor

SI

S2

S3

S3A

S4

S4A

S5

S6

S6A

S7

sg

S9

SIO

Number ofItems

3

1

3

3

3

1

1

1

1

2

1

1

1

fVfanagementFactor

Ml

M2

M3

M4

M5

M6

M7

M8

M9

Number ofItems

5

2

2

2

0

0

1

1

1

CultureFactor

Cl

C2

C3

C4

C4A

C5

C6

C6A

C7

C8

C9

CIO

cnC12

C1.3

Number ofItems

2

2

1

2

1

0

2

1

3

1

0

0

0

0

1

individual level. Hofstede (1984, pp. 43 and 55)admits,

"A between-cultures analysis had notbeen done at that time; first, because themain purpose of the survey operationwas organization development - that is,use within parts of the organization -within made the within-analysisobvious, and second, I must confess thatthe difference between within- andbetween-culture analysis had notoccurred to us at that time. If it had, wemight have come to a very differentselection of, in particular, the 'culture'survey items.. .From the earliest! surveysonward, it had been clear that questionsdealing with hierarchical relationshipsreceived systematically differentanswers in different countries."

From this point forward, while it is somewhatunclear in past writings, Hofstede derived hisfour separate cultural dimensions from the

"culture" factor as he saw themes emerge. Hederived these dimensions theoretically insteadof empirically. In other words, he examined the"culture" factor and made educated guesses atwhich items should make up each of his fourcultural dimensions.

The first dimension was power dislance.Hofstede noticed that the question, "Howfrequently are employees afraid to expressdisagreement with their managers?" wasreceiving similar answers within cultures (butnot between). He then decided to choose thisone core question iïs his entire power distancedimension (Hofstede 1984). Two additionalquestions were added based on ecologicalcorrelations and this formed the Power Distancedimension.

Next, according to Hofstede (1984, p. 55), "theuncertainty avoidance index was developed inan analogous way. I had an earlier theoreticalinterest in the phenomenon of stress which was



measured by the question 'How often do youfeel nervous or tense at work?'" Scores on thisquestion differed greater by country than byoccupation. Thus, Hofstede was able to deducethat a cultural dimension existed aroundpeople's differing reactions and exceptions touncertainty and anxiety.

A potentially rich source of data was alsoavailable in the "work goal importance"questions (Hofstede 1984). After normalizingthe data on these 14 questions, Hofstederealized that a structure emerged similar toMaslow's hierarchy of needs (Hofstede 1984).Through a review of the literature and a longprocess of analysis and deduction, Hofstedelater decided that these 14 questions measuredtwo constructs - Masculinity and Individualism.Subsequently, Hofstede subjected the abovementioned questions for all four factors, plusothers that had shown a reasonable amount ofstability over time to a factor analysis withorthogonal rotation. The final 32 itemsexplained 49 percent of the variance andbecame his initial set of questions to measurehis cultural dimensions.

Even though the masculinity and individualismconstructs were derived from the samevariables and even though the above factorloadings would empirically be three constructs,Hofstede took a different approach. For reasonswhich are not explained anywhere in theliterature, according to Hofstede (1984, p. 62),"Factor one represents an Individualism-lowPower Distance factor...I shall continue to treatthem as two dimensions because they areconceptually distinct...Factor 2 is a masculinityfactor... and factor three corresponds touncertainty avoidance" (Hofstede 1984, p. 62).These items can be found in Table 3. As can beseen, some items are included in overlappingdimensions. The obvious confusion of thismethodology, along with its limitations, isdiscussed in the following sections.

LIMITATIONS

Other than the obvious methodologicalproblems which stem from the above described

Marketing Management Journal, Fatl 2008

analyses, numerous other limitations exist withHofstede's research. While the overall findingsof Hofstede's research are extremely relevant totoday's cross-cultural studies, and the rigor ispossibly unmatched even today, majorconstraints exist with Hofstede's research. First,and through no fault of Hofstede, there is aquestion of time relevancy. Researchers havequestioned whether the dimensions developedfrom data collected between 1966 and 1973were artifacts of the period of analysis (e.g.,Baumgartel and Hill 1982; Warner 1981).Hofstede investigated the correlations betweenhis data and other variables like geographic,economic, demographic, and political nationalindicators. Over forty years have passed sincethe beginning of the study. Just a simple map ofthe world looks very different today than it didin 1966. While these correlations werebeneficial, they are not only out dated, but thecultures themselves have changed.

Second, Hofstede's research may suffer fromsampling problems. Several researchers haveargued that the constraints derived fromHofstede's research population of IBMemployees (e.g.. Graves 1986; Merker 1982;Triandis 1982). The use of employees from onecompany allowed Hofstede to reduce the othersources of variance and concentrate on culture.However, several criticisms have come fromthis fact. First, IBM employed mostly males atthe time of the survey. In the words of MiltonBennett (1996), "the differences between menand women is the greatest culture conflict ofall." More differences exist between men andwomen than from country to country, especiallyw h e n a n a l y z i n g t h i n g s l i k emasculinity/femininity, power distance, andindividualism/collectivism.

In lieu of the sampling issue, all the subjects inthe survey were from the same corporateculture. Additionally, although Hofstedesurveyed many countries, all subjects wereemployees of an American company.Additionally, most employees were from white-collar positions. Hofstede (1980) himselfdiscusses the problems of ethnocentrism thatexist in previous scales. As Hamden-Tumer and


TABLÉ 3Original Hofstede Items and Factor Groupings

Construct

Individualism andCollectivism Items

Items were used tomeasure bothIndividualism andMasculinity (Becauseitems bad factorloadings above O.S oneacb construct)

Masculinity andfemininity items

Power Distance items

ItemLabel inPresent

Stud

iciicioicllicl2icl3icl4ici 3ic2ic3ic4ic5ic6ic7ic8ic9

icandmfl

icandmf2

icandmD

mflmflOmfllmfl2

mn3

mn4mfl5mflómn7mfl 8mfl9mf2mf3mf4mf5mf6mf7mfSmf9

pdl

pdlOpdllpdl2pdl3pdl4pdl5pdl6pdl7

pd2pd3

Hofstede Item

Have good working conditions (good lighting, adequate work s]iace, an attractive office, etc.)?I would not support my work group if I felt they were wrong.If tbe group is slowing me down, it is better to leave and work a lone.It is better to work in a group tban alone.Groups make better decisions tban individuals.I prefer to be responsible for my own decisions.Contributing to the group is tbe most important aspect of work.Have considerable fi^edom to adopt your own approacb to tbe job?Have a job tbat leaves sufficient time for your personal or family life?Fully use your skills and abilities on tbe job?Have a job on wbicb tbere is a great deal of day-to-day learning?Competition among employees usually does more harm tban good.Decisions made by individuals are usually of bigber quality tbaii decisions made by groups.It is important to stick witb my work group, even tbrougb difficulties.My personal accomplishment is more important than group success.

Have challenging work to do; work from which you can get a pimsonal sense of accomplishment.

Having interesting work to do is just as important as having bigh earnings.

Most employees want to make a real contribution to tbe success of tbeir company.

Live in an area desirable to you and your family?Have tbe security that you will not be transferred to a less desirable job?Work in a congenial and friendly atmospbere?A corporation sbould bave a major responsibility for tbe healtb and welfare of its employees and tbeirimmediate families.A corporation sbould do as much as it can to help solve society's problems (poverty, discrimination,pollution, etc.).Most companies bave a genuine interest in the welfare of their umployees.The private life of an employee is properly a matter of direct concern to his company.It is important for me to have a job tbat provides opportunity for advancement.It is important tbat I outperform otbers in tbe company.It is important for me to have a job tbat provides an opportunity' for high earnings.It is important for me to work in a prestigious or successful company.Have an opportunity for high earnings?Work with people who cooperate well with one another?Have the security tbat you will be able to work for your company as long as you want to?Have an opportunity for advancement to bigber-level jobs?Have a good working relationship with your manager?Ciet the personal recognition you deserve when you do a good job?Have a job tbat allows you to make a real contribution to tbe success of your company?Work in a company tbat is regarded in your coimtry as successlul?

Employees lose respect for a manager wbo asks tbem for tbeir iidvice before be makes a fmaldecision.For getting abead in industry, knowing influential people is usually more important tban ability.Even if an employee feels that he deserves a salary increase, he should not ask his manager for it.My superiors should make most decisions without consulting n le.It is improper to disagree with one's supervisor.I would never argue with my supervisor.I believe that those superiors who ask opinions too often of subordinates are weak or incompetent.I believe tbat superiors are entitled to special privileges.77iis question asks the respondent to circle his preferred manager type among three choices, from themost consultative to the least consultative.Employees sbould participate in tbe decisions made by managnment.Company rules sbould not be broken; even wben tbe employee thinks it is in the company's bestinterests.


A Re-Inquiry of Hofstede's Cultural... . Orr and Häuser

TABLE 3{continued)

Construct

PowerDistance Items(continued)

UncertaintyAvoidance items

Item Labelin present

study

pd4pd5

pd6pd7

pd8

pd9

ual

ua2ua3ua4ua5ua6ua7

Hofstede item

Employees should never express disagreement with their managers.Employees should always be told very clearly their duties and responsibilities, and how to performtheir jobs.Most employees have an inherent dislike of work and will avoid it if they can.A good manager gives his employees detailed and complete instructions as to how they should dotheir jobs; he does not merely give general directions and depend on them to work out the details.In general, the better managers in a company are those who have been with the company the longesttime.There are few qualities in a man more admirable than dedication and loyalty to his company.

A good manager does not get too involved in the details of an employee's job; rather, these details areleft to the employee.Staying with one company for a ling time is usually the best way to get ahead in business.A large corporation is generally a more desirable place to work than a small company.Companies should not change their policies and practices very often.It is important for me to work for a company that provides high employment stability.Clear and detailed rules/regulations are needed so workers know what is expected of them.It is better to work in a well-defined job where the requirements and procedures are clear.

Trompenaars (1997) noted, they doubt thatAmerican IBM managers serving in foreigncountries are much different than AmericanIBM managers in America.

Inclusive with the sampling issue is the matterof the original sample size. While the sampleeventually grew to quite some size, the originalconstructs were derived from very few workers.According to Hair and colleagues (1998), theminimum sample size is five observations pervariable to be analyzed. However, tenobservations per variable are better, and someeven recommend 20 per variable. Therefore,Hofstede's results were sample specific andthey took advantage of random correlations.Some samples were as low as 231 for 146items.

Hamden-Tumer and Trompenaars (1997) bringup another important criticism of Hofstede'swork. "Are cultural categories linear andexclusive?" Hamden-Tumer and Trompenaars(1997) do not feel that if you are anindividualist you cannot be a coUectivist.Perhaps some people tend to be veryindividualistic at work, but family oriented andcollectivistic at home.


Another criticism is that culture cannot be bestexpressed in a mathematical language, asHofstede does. This point can be bestsummarized by a story given by Hamden-Tumer and Trompenaars (1997). A leamedresearcher diced a piece of cheese with akitchen gadget and then wrote a leameddissertation on the cubic nature of cheese. Weget out of factor analysis what we put into it,nothing more, and nothing less. As the previoussection described, Hofstede's determinationswere haphazard at best.

Another problem with Hofstede's work is thatthe study did not begin as a cultural study. Itinitially began as a work satisfaction study.Hofstede was a brilliant researcher who noticedthe dimensions as they developed. However,the original survey was not designed for itsfinal purpose. The survey was refined andchanged several times to make the necessaryadjustments. Thus, the dimensions were derivedempirically, rather than theoretically.

Finally, Hofstede (1980) specifically identifiedthe ecological fallacy that exists with his work.The ecological fallacy can be defined as"confusion between within-system and

8

A Re-Inquiry of Hofstede's Cultural.. . . Orr and Häuser

ecological correlations." Similarly, he readilyadmits that within-culture variations can be asgreat as if not greater than between-culturevariations. This is an important observation tomake. While it is helpful to understand that themajority of Chinese citizens tend to be verycollective, marketers sell to the individual,mangers recruit, train and hire the individual,and psychologists, economists, and all socialscientists are concerned with the individual andgroups existing in their society.

The above mentioned limitations have beenexamined and identified in numerousreplication studies over the last four decades.Sendergaard (1994) located 61 replications ofHofstede's research. Full confinhation ofHofstede's dimensions was found in only fourstudies (Hoppe 1990; Shackelton and Ali1990). Partial confirmation was found inanother fifteen studies (e.g., Ashkanini 1984;Chow et al. 1991; Forss 1989; Huo and Randall1991; Lowe 1994; Maldonado 1983; Pooyan1984; Westwood and Everett 1987; Yeh 1988).Lowe's (1994) study is particularly interestingbecause he used IBM employees from HongKong and the United Kingdom for his sample.Lowe was not able to find differences betweenthe two countries for Hofstede's uncertaintyavoidance dimension (S0ndergaard 1994).

After Sondergaard's (1994) study, other authorshave critiqued and replicated Hofstede's workand applied his dimensions to various contexts(e.g., Fernandez, Carlson, Stepina, andNicholson 1997; Kelleher 2000; Marshall 1997;Naumov and Puffer 2000; Robertson andHofftnan 2000; Smith, Dugan, Peterson, andLeung 1998; Sopachitwattana 2000;Trompenaars 1993; Trompenaars 1997; VanOudenhoven 2001; Van Oudenhoven,Mechelse, and de Dreu 1998; Verbeke 2000;Yeh and Lawrence 1995). Most of these studieshave come to very similar conclusions as theones prior to 1994. No known rigorous studyusing Hofstede's exact instrument has foundcomplete confirmation to Hofstede's work. Infact, Trompenaars (1993) examined Hofstede'sdimensions and arrived at his own dimensions,which, according to Trompenaars, overcome

the difficulties with Hofstede's research.Trompenaars' book (1993) has become a harshdebate between Hofstede and himself (for fulldetails, see, Hamden-Tumer and Trompenaars1997; Hofstede 1996; Hofstede 1997).

THE METHODOLOGYOF THE, RE-INQUIRY

Even with all the rejüication studies that exist inthe literature, very few have been exactreplications, using Hofstede's actual, originalitems. In fact, prior to 2001, Hofstede did notmake these original items readily available.Thus, the purpose of this study was to producea very similar replication study. A further aimof this study was to analyze the dimensionsusing multiple statistical techniques in order toexamine the items and constructs as thoroughlyas possible. In order to start this process,Hofstede's original instrument was obtainedfrom Culture's Consequences. The survey wasadministered in two different forms. The firststudy was an attemjit to assess face validity. Inthe second study, convergent and discriminantvalidity was assessed. Finally, the data is testedthrough confirmatory factor analysis utilizingstructural equations tnodeling.

In order to ascertain the effects of theinstrument on a non homogeneous sample (e.g.,samples from more than one company) anumber of different samples were used. Thiswas also done to increase the probability ofvariance from other sources, such as gender,while minimizing the confounding effects ofsuch factors as occupational status (e.g., whitecollar) within one over-arching corporateculture environment.

Study One

In study one, Hofstede's instrument, wilh nomodifications, was administered to two samples(refer to Table 3). The first sample included123 undergraduate students and the secondsample included 65 graduate students, and thethird sample coniained 13 marketing andmiinagement faculty. The undergraduate classtook the survey after a whole class period

Marketing Management Joumat, Fall 2008


dedicated to teaching Hofstede's culturaldimensions. All samples were given the surveywith thorough definitions of each of Hofstede'sfour cultural dimensions at the top of theinstrument. The subjects were then instructed toascertain which dimension each survey itemwas attempting to measure by circling IC( i n d i v i d u a l i s m / c o l l e c t i v i s m ) , MF(masculinity/femininity), UA (uncertaintyavoidance), or PD (power distance) after eachitem.

As can be seen in Table 4, the task proved to bequite difficult. The graduate students performedbetter than the faculty, followed by theundergraduate students, ascertaining correctly64.62 percent, 44.25 percent, and 32.25 percentof the time respectively. For a face validityassessment, the percentages should be muchhigher, whether the sample is common workersor especially trained researchers and academicssuch as those used in this study. Theindividualism items were easiest to identifywhile the masculinity items were the hardest toclassify.

Study Two

In study two, Hofstede's exact survey with nomodifications was given to three differentsamples. Two out of the three samples were inAmerica. The first sample consisted of graduate

students at a large mid-south university. Of the161 respondents, 58 percent were America and41.6 percent were Far Eastern. Respondentswere required to have full time work experiencein order to participate in the study. This criteriawas added to the survey so that all respondentscould identify with the work-related questionsthat appear on Hofstede's instrument.Respondents had work experience ranging from1 to 40 years and were between 22 and 59 yearsof age. A second sample of non-student adultswas taken in America (N = 233). Demographiccharacteristics of this sample were very similarto sample one, except 97.4 percent of thesample was American. Finally, a third samplewas gathered from non-student adults inAustralia (N = 210). Due to the widespreadcultural differences that exist in Australia, thissample was extremely diverse. The samplecontained people with varying nationalitiesrepresenting a total of 21 countries.

Data were analyzed in many different ways inorder to demonstrate that the factor structuredoes not hold in any circumstance. Reliabilitieswere assessed for each of the four factors withthe three combined samples and with thesamples separately. Likewise, factor analyzeswere run with each sample separately andcombined. Exploratory factor analyses wereused to remain consistent to Hofstede's originalmethodology and many different techniqueswere utilized. Varimax and oblique rotations

TABLE 4Face Validity Assessments: Average Number of Times

Ascertained Correctly Across Dimension

Masculinity items

Individualism items

Power Distance items

Uncertainty avoidance items

Average

Facultysample

28%

55%

41%

53%

44.25%

Graduatesample

40.73%

78.24%

60.94%

78.57%

64.62%

Undergraduatesample

15%

46%

33%

35%

32.25%

Average of allthree samples

27.91%

59.75%

44.98%

55.52%

47.04%

Marketing Management Journal, Fall 2008 10

A Re-Inquiry of Hofstede's Cultural . . . . Orr and Häuser

were tried to remain true to Hofstede. Likewise,factors were extracted by examiningeigenvalues greater than one and by "forcing"the solution to only four factors. Additionally,even though Hofstede did not use confirmatoryfactor analysis, this study attempts to do so. Allanalyses will be discussed subsequently.

Reliabilities

Reliabilities were examined subsequent toperforming the factor analyses. The results arepresented in Table 5. Some reliabilities wereextremely low (.3405) and some are relativelyhigh (.8131). Curiously, the data shows noconsistent pattern across the samples. In otherwords, masculinity has the only adequatereliabilities in sample two, but is the secondlowest in sample one, both of which wereAmerican samples.

Exploratory Factor Analyses

As mentioned, factor analyses were examinedin many possible ways. Data are presented foreach sample separately and then are presentedby combined samples. As can be seen inTable 6, in all samples, when the number ofcomponents is not forced, the instrumentexplained around 70 percent of the variance.However, rotation could not account forcoverage many times, and when it could, thenumber of factors ranged between 15 and 20.When the data were forced into four

components, explainable variance dropped toaround 30 percent.

These statistics seem discouraging, but it iseven more discouraging to analyze what itemsare loading with which construct. In allanalyses, items had no pattern as to whichconstruct they loaded with. This is the case forall samples separately and together, withvarimax and oblique rotation, and with Eigenvalues greater than one or forced factors. Dueto space constraints, all 32 factor structurescannot be presented. Explanation orclassification of each component is notpossible. The components' items have nopattern or similarities. There was absolutely notheoretical and empirical structure whichemerged.

Confirmatory Fador Analysis

Next, structural equations modeling was used toanalyze the data because of its ability to be usedas a confirmatory technique, instead of as anexploratory technique, even though Hofstededid not originally employ this technique. Sincethe relationships and constructs had alreadybeen established by Hofstede, the model justneeded to be analyzed to assess the adequacy ofthe model. The three previous mentionedsamples were utilized to test the model.

The results of the model showed an inadequatefit. The chi-squared/df ratio was equal to 5.345

TABLESReUabilities (Study 2)

Masculinity

Individualism

Power Distance

UncertaintyAvoidance

Sample 1(US graduate students)

.5254

.4816

.6974

.6330

Sample 2 (US non-student adults)

.7593

.3703

.3510

.3405

Sample 3(Australian)

.8131

.6527

.7003

.6845

All samplestogether

.7574

.4920

.7417

.5655

11 Marketing Management Journal, Fall 2008


TABLEÓExploratory Factor Analyses (Study 2)

Varimaxrotation/ Eigenvalues greaterthan one

Oblique rotation/Eigen valuesgreater than one

Varimaxrotation/ Four"forced"components

Oblique rotation/Four "forced"components

Sarnple 1(US graduate students)

Numberof factors

19

Rotationcould notconverge

in 100iterations

4

4

Varianceexplained

70%

70%

31.7%

30%

Sample 2(US non-student adults)

Number offactors

Rotationcould notconverge

in 100iterations

20

4

4

Varianceexplained

69.6%

69.6%

30%

30%

Sample 3(Australian)

Number offactors

Rotationcould not

converge in100

iterations

Rotationcould not

converge in100

iterations

4

4

Varianceexplained

69.2%

69.2%

37%

37%

All samples together

Number offactors

15

15

4

4

Varianceexplained

60%

60%

36%

36%

(8493 with 1589 d.f.). Ratios under threeindicate an acceptable fit (Carmines and Mclver1981). The Root Mean Square Error ofapproximation (RMSEA) by Browne andCudeck (1993), which is a goodness of fitmeasure that accounts for model complexity,was 0.85. Browne and Cudeck (1993) state thatRMSEA values of about .05 or less indicate aclose fit of a model in relation to the degree offi"eedom. Likewise, the normed fit index(Bentler and Bonnett 1980) was 0.887, whichshould be above 0.90, which indicates andacceptable level of fit (Hair, Anderson, Tathamand Black 1998). Results of these analyses arepresented in Tables 7 and 8.

The reliabilities and the variances explained byeach of the latent contructs are presented inTable 8. The highest reliability was for themasculinity construct (a = 0.73), followed bypower distance (a = 0.58), then by Uncertaintyavoidance (a = 0.40), and then withindividualism having the lowest reliability (a =0.33). According to Hair et al. (1998), theindicator reliabilities should exceed .50, whichroughly corresponds to a standardized loadingof 0.70.

The results of the variance explained by eachconstruct had even worse results. The highestvariance extracted was for the uncertaintyavoidance construct (16.54 percent), followedby power distance (10.33 percent), then byindividualism (10.33 percent), and then bymasculinity having the lowest total varianceextracted (7.97 percent). According to Hair etal. (1998), guidelines suggest that the varianceextracted value should exceed .50 for aconstruct. None of the constructs had valueabove this percentage.

In regards to factor loadings, some items havehigh factor loadings and are significant.However, a larger number have very lowloadings and are not significant. Amazingly, notone single power distance item is significant.While masculinity, individualism, anduncertainty avoidance may be adequateconstructs (by these criteria only) once a fewimnecessary items are eliminated, powerdistance has clear empirical problems.

Additionally the correlations among latentconstructs are provided in Table 9. The moststriking correlation is that of masculinity and

Marketing Management Joumal, Fall 2008 12

A Re-Inquiry of Hofstede's Cultural.., Orr and Häuser

TABLE 7CFA Results

Masculinity Individualism Power Distance Uncertainty Avoidance T-valueReliability = 0.73 Reliability = 0.3 Reliability = 0.58 Reliability = 0.40

Variance Extracted = Variance Extracted = Variance Extracted = Variance Extracted =7.97% 9.15% 10.33% 16.54%

mf2_l

mf3_l

mf4_l

mf6_mf7_mf8_mf9_mflOmm

mfl4mfl5

idic2ic3.ic4.ic5.ic6ic7.ic8.ic9_iclO_icll_icl2_icl3_icl4_icl5_pdl_pd2_pd3_pd4_pd5_pd6_pd7_lpd8_l

pd9_lpdIO_pdll_pdl2_pdl3_pdl4_pdl5_pdl6_pdl7_

0.55

0.50

0.66

0.64

0.580.660.620.640.570.660.70

-0.10-0.020.04

-0.08-0.04-0.15-0.13-0.05

0.640.550.540.630.580.08

-0.03-0.050.050.02

-0.01-0.04-0.09-0.07-0.22

0.05-0.090.300.750.190.52

-0.100.330.17

-0.070.290.550.460.790.320.62

-0.73

12.212***

10.099***

12.224***

11.968***

11.241***12.190***11.652***11.998***11.031***12.216***12.682***

-2.282*-0.5030.832

-1.878-0.991

-3.455***-2.97**

-1.10113.368***11.858***11.767***13.366***12.329***

1.825-0.757-1.2621.1500.418

-0.245-0.906

-2.043*-1.728

-5.047***0.545

-0.9981.12

1.1361.1011.133

•1.0131.126

1.091.0.9091.1211.1331.1321.1361.1251.134

-1.135



TABLE 7CFA Results (continued)

ua7_

ua6_

ua5

ua4_

ua3

ua2

ual_

MasculinityReliability = 0.73

Variance Extracted =7.97%

IndividualismReliability = 0.3


Power DistanceReliability = 0.58


Uncertainty AvoidanceReliability = 0.40


0.82

0.11

0.77

0.45

-0.15

-0.01

-0.08

T-value

16.436***

2.493*

15.673***

10.032***

-3.386***

-0.105

-1.862

*p < .05**P<.01***p < .001Items listed in Table 4.

TABLE 8CFA Results: Correlations Among Latent Constructs

Masculinity

IndividualismPower Distance

Uncertainty Avoidance

Masculinity

1.000

0.987

-0.218

-0.012

Individualism

1.000

-0.093

-0.059

Power Distance

1.000

-0.651

UncertaintyAvoidance

1.000

individualism (0.987). Perhaps this is becauseHofstede used the same items to measure eachconstruct. Hofstede states, "...reversing thesign of the scores (for the items forindividualism), I have called this dimension'Masculinity'" (Hofstede 1984, p 189). Itemsthat load negatively on a construct are not aseparate construct, just the opposite "pole" ofthat construct. Power distance and Uncertaintyavoidance also have a high correlation betweenthem (-0.651). Many of the power distanceitems relate to the supervisor's responsibility toestablish clear rules and regulations (e.g.,"Employees should always be told very clearlytheir duties and responsibilities, and how toperform their jobs" and "A good manager giveshis employees detailed and completeinstructions as to how they should do their jobs;he does not merely give general directions anddepend on them to work out the details").Conceptually, one can see how these itemscorrelation with uncertainty avoidance. People

high in uncertainty avoidance want clear rulesand regulations so that there is less uncertaintyto deal with.

DISCUSSION

This study has attempted to empirically assessthe validity of Hofstede's cultural dimensions.Substantial questions have arisen in thisanalysis as to the reliability and validity ofHofstede's methodology and instrumentation.In fact, the analysis discussed in this re-inquirydoes seem to reify many of the limitations ofHofstede's work discussed earlier.

First, the samples used in this analysis werecomprised of diverse individuals who were notpart of an over-arching collectivity, namelyIBM employees. Thus the impact of acorporate culture was not in play here. Notonly does this minimize the corporate-wide


A R e - I n q u i r y o f H o f s t e d e ' s C u l t u r a l . . . . Orr and Häuser

socialization effects as to how one should thinkand act within that corporate culture, itminimizes the possibility of socially desirablerespondents by the respondents.

Second, limitations around the lack of themutual exclusivity of the dimensions alsosurfaced in this re-inquiry. This analysis foundsignificant overlap within and across many ofthe constructs. Thus, with the samples usedherein, there were no clearly identifiable factorssupporting the instrumentation and, to a largedegree, the methodology in its current state.

Third, and probably most importantly, theanalysis questions the relevance of the originaldimensions and their meaning to 21^' centurybusinesses and individuals. What do thedimensions mean to individuals within andacross different cultures? What about sub-cultural differences that exist in many countriesand regions? Most importantly, what effect dotraditional social institutions have on thedimensions defined by Hofstede? For example,strong religious dogma and practices in aculture will most likely strongly impact anindividual's perception of individualism as wellas culturally sanctioned definitions ofmasculine and feminine roles. Similarly, legal,economic, and educational institutions withinthe given social structure will dramaticallyinfluence how one responds to Hofstede'sdimensions.

This analysis also tends to support a number ofHofstede's critics as to the applicability of thefour (or five) dimensions. In general, thedimensions only attempt to measure culturaldifferences at the individual level and aretherefore psychologically reductionistic. Cross-cultural analysis requires an understanding theimpact of the socialization and othersociological factors that brought these about.As a matter of fact, Hofstede has longcontended that an ecological fallacy iscontained in the cultural dimensions. Ifanything, the current analysis suggests that areverse ecological fallacy may be the casewhere individual characteristics are being

assigned to an entire group or in this caseculture. This analysis discussed supports theneed to understand differences at the societallevel. Currently, this lack of understanding ofthe causality between the individual level andthe socio-structural level precludes any clearindication of what is actually causingdifferences within and across Hofstede'sdimensions.

Finally, the impact of the Internet and relativelyseamless global communications on a society'scultural stance cannot be over-estimated in theearly 21^ centuiy. Short of extremegovernmental contol, it is nearly impossiblefor individuals in any culture to not have accessto and be infiuencod by information on othercultures, attitudes, and behaviors. Futureresearch on Hofstede's cultural dimensionsmust investigate the impact of globalcommunications on cultural dimensions andindividual responses to them.

CONCLUSIONS

Hofstede's seminal work has been thebenchmark for cultural analysis for the lastthree decades. However, it has been subject tocriticism on both tlie theoretical and empiricallevels. The intent of this investigation was totest Hofstede's constructs with a non-homogenous population (i.e., individuals thatwere not from one company only) in order toascertain the validity and reliability of themeasures. In order to do so a number ofsamples drawn fi-om American, Far EasternAsians, and Australians were used. Analyseswere performed into the face validity,convergent, and discriminant validity ofHofstede's constructs. Likewise, Hofstede'sfactors were subjected to exploratory andconfirmatory factor analysis where theyperformed poorly. This investigation hasconcluded that Hofstede's factors overlapsignificantly and do not share a common factorstructure within or between cultures.

While it is outside the scope of thisinvestigation, it appears that Hofstede's

IS Marketing Management Journal, Fall 2008


theoretical constructs need to be thoroughly re-examined within the context of early 21^'century cross-cultural attitudes and patterns ofbehavior. Cross-cultural relationships (positiveand negative) have changed dramatically overthe past quarter of a century, be they political,economic, or from a business perspective.Worldwide political systems, such ascommunism, have dramatically lost theirinfluence since Hofstede first posited hiscultural dimensions. Free market economieshave taken a foothold (to varying degrees) inmany cultures while businesses have becomemore global in their reach and influence.

At the same time, the changes mentioned abovehave exacerbated within society changes inmany regions of the world. As political andeconomic systems decayed, long constrainedcultural and sub-cultural differences have re-emerged, in a number of cases, to the pointwhere the country has been divided into anumber of smaller cultural or historically"tribal" based enclaves. With this in mind,Hofstede's original dimensions may beinaccurate, or at the least, outdated in definingcontemporary cultural differences be theywithin or across different cultures.

With regard to research implications, empiricalestablishment of convergent, discriminant, andnomological validity for the culturaldimensions are of first importance. If theconstructs are not defined empirically, thenthey cannot be measured. Likewise, ifHofs tede ' s dimensions cannot beoperationalized, then they cannot be correlatedwith other concepts or used in other studies tohave practical significance. The study of cross-cultural values is simply too important in thistime of globalization. Hofstede's dimensions ,are not reasonable empirically. Although theremay be a conceptual gold mine imdemeath itall, a theory is worthless to investigators if itcannot be operationalized.

In conclusion, the purpose of this re-inquirywas to re-examine Hofstede's originalmethodology to test the validity with a numberof diverse samples. While this study found

both validity and reliability issues with theoriginal constructs and instrumentation, it is notour intent to denigrate Hofstede's originalconceptualization. Instead, we recommend thatadditional research be undertaken to build onHofstede's cross-cultural dimensions to betteradapt them to the 21*' century globalenvironment. As such, each construct shouldbe thoroughly re-examined and both betheoretically and operationally redefined withincontemporary cross-cultural and businessenvironments.

Hofstede's work has contributed significantlyto the foundations of cross-cultural analysis andunderstanding. But like any good model, itneeds to constantly be re-examined, re-defined,and adapted to the current envirorunent.Therefore, ñiture studies should attempt tobuild on, strengthen, and adapt what has beenlearned from Hofstede's seminal work.Perhaps as we have moved well into this newmillennium very different cross-cultural values,attitudes, and behaviors exist.

REFERENCES

Ashkanini, Mohammed G. A. (1984), A Cross-Cultural Perspective on Work Related Value,Unpublished doctoral dissertation.International University, San Diego.

Baehr, Melany E. (1954), "A Factorial Study ofthe SRA Employee Inventory," PersonnelPsychology, 7, 319-336.

Baumgartel, Howard and Thomas Hill (1982),"Geert Hofstede: Culture's Consequences:International Differences in Work-RelatedValues," Personal Psychology, 35(1),192-196.

Bennett, Milton (1996), Better Together ThanA-P-A-R-T: Intercultural Communication: AnOverview, [videorecording]. Newtonville,Mass.: Intercultural Resource Corporation.

Bentler, Peter M. and Douglas G. Bonnett(1980), "Significance Tests and Goodness ofFit in the Analysis of Covariance Structures,"Psychological Bulletin, 88, 588-606.



Berry, John W. (1997), "Preface." In Berry, J.W., Poortinga, Y. H. and Pandey, J. (EA),Handbook of Cross-Cultural Psychology,Volume 1: Theory and Method. NeedhamHeights, MA: Allyn and Bacon.

Browne, Michael W. and Robert Cudeck(1993), "Alternative Ways of AssessingModel Fit," In Testing Structural EquationModels, Kenneth A. Bollen and Scott J.Long, eds. Newbury Park, CA: Sage.

Carmines, Edward G. and John P. Mclver(1981), "Analyzing Models with UnobservedVariables," In Social Measurement, CurrentIssues, George W. Bohmstedt and Edgar F.Borgatta, (Ed), Beverly Hills: Sage.

Chinese Culture Connection (1987), "ChineseValues and the Search for Culture-FreeDimensions of Culture," Journal of Cross-Cultural Psychology, 18, 143-164.

Chow, Chee W., Michael D. Shields and YokeKai Chan (1991), "The Effects ofManagement Controls and National Cultureon Management Performance," Accounting,Organizations, and Society, 209-226.

De Cieri, Helen and Peter J. Dowling (1995),"Cross-Cultural Issues in OrganizationalBehavior," In Cooper, C. L. and D. M.Rousseau (Ed), Trends in OrganizationalBehavior. West Sussex, England: John Wileyand Sons.

Etzioni, Amitai (1975), A ComparativeAnalysis of Complex Organizations (Reviseded.). New York: The Free Press.

Fernandez, Denise Rotondo, Dawn S. Carlson,Llee P. Stepina and Joel D. Nicholson,(1997), "Hofstede's Country Classification25 Years Later," Journal of SocialPsychology, \37(^l), 43-54.

Forss, Kim (1989), Comparative Managementin Public Administration: A Pilot Study ofValues and Attitudes in Two Organizations,The Central Bureau of Statistics, TheNational Institute for Civil Service Trainingand Development and the National TaxBoard.

French, John R. P. and Bertram Raven (1959),"The Bases of Social Power," In D.Cartwright (Ed), Studies in Social Power.Ann Arbor Michigan: ISR.

Galton, Francis [1883] (1928). Inquiries intoHuman Faculty and its Development.London: Dent.

Graves, Desmond (1986), Corporate CultureDiagnosis and Change: Auditing andChanging the Culture of Organizations.London: Frances I'inter.

Hair, Joseph F., Ronald E. Anderson, Rolph L.Tathan and William C. Black (1998),Multivariate Data Analysis: (5* ed.). UpperSaddle River, NJ: Prentice Hall.

Hamden-Turner, Charles and FonsTrompenaars (1997), "Response to GeertHofstede," International Journal ofIntercultural Relations, 21(1) 149-159.

Herzberg, Fredrick, Bernard Mausner, O.Peterson and Dora F. Capwell (1957), JobAttitudes: Review of Research and Opinion.Pittsburgh, PA: Psychological Service ofPittsburg.

Hinrichs, J. R. (19(í8), "A Replicated Study ofJob Satisfaction Dimensions," PersonnelPsychology. 21(4), 479-503.

Hofstede, Geert, A. L. Kraut and S. H.Simonetti (1971), Development of a CoteAttitude Study Questionnaire forInternational Use, Corte Madera, CA: SelectPress.

(1975), The Stability of Attitude SurveyQuestions: In Particular Those Dealing withWork Goals, Unpublished paper, Brussels:European Institute for Advanced Studies inManagement.

(1980), Culture's Consequences:International Differences in Work RelatedValues, Thousand Oaks, CA: SagePublications, Inc.

(1983), "National Cultures in FourDimensions: A Research-Based Theory ofCultural Differences Among Nations,"International Studies of Management andOrganization, 13,46-75.

(1988), "The Confucius Connection: FromCultural Roots to Economic Growth,"Organizational Dynamics, 16,4-22.



(1994), Forward in U. Kim, Harry C.Triandis, C. Kagitqibasi, Sang-Chin and G.Yoon, (Ed), Individualism and Collectivism:Theory, Method and Applications, (pp. ix-xiii). Thousand Oaks, CA: Sage Publications,Inc.

(1996), "Riding the Waves of Commerce:A Test of Trompenaars' 'Model' of NationalCultural Differences," International Journalof Intercultural Relations, 20(2), 189-198.

(1997), "Riding the Waves: A Rejoinder,"International Journal of InterculturalRelations, 21(2), 287-290.

(2001), Culture's Consequences:International Differences in Work RelatedValues, 2001 edition. Thousand Oaks, CA:Sage Publications, Inc.

Hoppe, Michael H. (1990), "Testing theConstruct Validity of Geert Hofstede'sMasculinity-Femininity Dimension on Elitesfrom Nineteen Countries," Paper for XlthInternational Congress of Cross-CulturalPsychology, Liege.

Hsu, Francis L. K. (1971), "PsychosocialHomeostasis and Jen: Conceptual Tools forAdvancing Psychological Anthropology,"American Anthropologist, 11, 23-44.

Huo, Pau Y. and Donna M. Randall, (1991),"Exploring Subcultural Differences inHofstede's Value Survey: The Case of theChinese," Asia Pacific Journal ofManagement, 892, 159-173.

Institute for Research on InterculturalCooperation (2001), "Europeans in the 2000Social Science Citation Index,"http://cwis.kub.nl/~fsw 2/iric/index2.htm.

International Association of Cross-CulturalPsychology (2007), "New IACCP HonoraryFellows," http://www.iaccp.org/news_page.php?thisID=441 andnewstype=news.

Jahoda, G. and B. Krewer (1997), "History ofCross-Cultural and Cultural Psychology," InJ. W. Berry, Y. H. Poortinga, and J. Pandey,(Ed), Handbook of Cross-CulturalPsychology, Volume 1: Theory and Method.Needham Heights, MA: Allyn and Bacon.

Kahle, Lynn R. (1983), Social Values andSocial Change: Adaptation to Life inAmerica, New York: Praeger.

Kelleher, B. R. (2000), "A Cross-CulturalStudy of Work-Related Values: AnInvestigation Across Three Countries UsingHofstede's Cultural Indices," DissertationAbstracts International Section A:Humanities and Social Sciences, 60(7-A),2555.

Lowe, S. (1994), "Hermes Revisited," Paper forthe ISPAB Conference, Bangkok.

Maldonado, E. M. (1983), "Mexican-AmericanParticipation in Law Enforcement inCalifornia: Law Enforcement the Impact ofCultural Related Value Differences,"Unpublished doctoral dissertation, ClaremontGraduate School, Claremont.

Marshall, R. (1997), "Variances in Levels ofIndividualism across Two Cultures," Journalof Cross-Cultural Psychology, 28(4),490-496.

Martinsons, M. G. and R. I. Westwood (1997),"Management Information Systems in theChinese Business Culture: An ExplanatoryTheory," Information and Management, 32(5), 215-229.

Merker, S. (1982), "Geert Hofstede: 'Culture'sConsequences'," Behavioral Science. 27 (2),195-197.

Naumov, Alexander I. and Shelia M. Puffer(2000), "Measuring Russiem Culture UsingHofs tede ' s Dimens ions ," AppliedPsychology: An International Review, 49(4),709-718.

Parsons, Talcott and Edward A. Shils (1951),Toward a General Theory of Actions,Cambridge, Massachusetts: HarvardUniversity Press.

Pooyan. Abdullah (1984), "Acculturation,Acculturative Stress, and their Relationshipsto Both Work and Non-Work Outcomes,"Unpublished doctoral dissertation, Dallas:University of Texas.

Redding, S. Gordon (1994), "ComparativeManagement Theory: Jungle, Zoo, or FossilBed?" Organization Studies, 15(3), 232-359.

Robertson, Christopher J. and James J.Hof&nan (2000), "How Different Are We?An Investigation of Confucian Values in theUnited States," Journal of Managerial Issues,12, 34-47.



Rokeach, Milton (1973), The Nature of HumanValues, New York: The Free Press.

S0ndergaard, Mikael (1994), "Research Note:Hofstede's Consequences: A Study ofReviews, Citations and Replications,"Organization Studies, 15(3), 447-456.

Schwartz, Shalom H. and W. Bilsky (1987),"Toward a Universal Psychological Structureof Human Values," Journal of Personalityand Social Psychology, 53(3), 550-562.

Shackelton, Viv J. and Abbas H. Ali (1990),"Work Related Values of Managers: A Testof Hofstede's Model," Journal of Cross-Cultural Psychology, 21(1), 109-118.

Smith, Peter B., Shaun Dugan, Ark Petersonand Wok Leung (1998), "Individualism:Collectivism and the Handling ofDisagreement. A 23 Country Study,"International Journal of InterculturalRelations, 22(3), 351-361.

Sopachitwattana, P. (2000), "An Examinationof the Effects of Participative Managementand Cultural Values on OrganizationalCommitment for American and ThaiManagers," Dissertation AbstractsInternational Section A: Humanities andSocial Sciences, 61(3), 1075.

T0nnies, Ferdinand (1963) [1887], Communityand Society, New York: Harper and Row.

Triandis, Harry C. (1982), "Culture'sConsequences," Human Organization, 41 (1),86-90.

Trompenaars, Fons (1993), Riding the Waves ofCulture: Understanding Diversity in GlobalBusiness, New York: Irwin.

Tucker, Ledyard R. and Charles Lewis (1973),"The Reliability Coefficient for MaximumLikelihood Factor Analysis," Psychometrika,38, 1-10.

Van Oudenhoven, Jan Pieter (2001), "DoOrganizations Reflect National Cultures? A10-Nation Study," International Journal ofIntercultural Relations. 21(1), 149-159.

, L. Mechelse, and C. K. W. de Dreu(1998), "Managerial Conflict Management inFive European Countries: The Importance ofPower Distance, Uncertainty Avoidance, andMasculinity," Applied Psychology: AnInternational Review, 47(3), 439-455.

Verbeke, Willem (2000), "A Revision ofHofstede et al.'s (1990) OrganizationalPractices Scale," Journal of OrganizationalBehavior, 21(5), 587-602.

Vroom, Victor H. (1964), Work andMotivation, New York, Wiley.

Warner, Malcolm (1981), "Culture'sConsequences," Journal of GeneralManagement. 1 (1), 75-78.

Westwood, Robert G. and James E. Everett(1987), "Culture's Consequences: AMethodology for Comparative ManagementStudies in Southeast Asia," Asia PacificJournal of Management, 4 (3), 187-202.

Wherry, Robert J. (1954), "An Orthogonal Re-Rotation of the Baehr and Ash Studies of theSRA Employees Inventory," PersonnelPsychology, 1, 365-380.

Yeh, Ryh-Song (1988), "Values of American,Japanese, and Taiwanese Managers inTaiwan: A Test of Hofstede's Framework,"48* Aimual Meeting of the Academy ofManagement, Anaheim. Academy ofManagement Best papers Proceedings 1988.

Yeh, Ryh-Song and John J. Lawrence (1995),"Individualism aiid Confucian Djmamism: ANote on Hofstede's Cultural Root toEconomic Growth," Journal of InternationalBusiness Studies, 26(3), 655-670.

Yoo, Boonghee aiid Naveen Donthu (2002),"Books in Review," Journal of MarketingResearch, (39) August, 388-389.


A RE-INQUIRY OF HOFSTEDE'S CULTURAL DIMENSIONS: A CALL …

Documents

Transcript of A RE-INQUIRY OF HOFSTEDE'S CULTURAL DIMENSIONS: A CALL …