Affective and continuance commitment to the organization: Evaluation of measures and analysis of...

International Journal of Intercultural Relations

26 (2002) 525–550

Affective and continuance commitmentto the organization: test of an integrated model

in the Turkish context

S. Arzu Wasti*

Graduate School of Management, Sabanci University, Orhanli Tuzla, 81474 Istanbul, Turkey

Abstract

The present study investigated organizational commitment in Turkey, a predominantly

collectivist society. A model of antecedents and consequences of organizational commitment

was tested, where commitment was conceptualized as composed of two dimensions, affective

and continuance. Affective commitment was hypothesized to develop from positive work

experiences and to predict desirable outcomes. Continuance commitment, on the other hand,

was argued to be culture-bound. In line with Becker (Am. J. Sociol. 66 (1960) 32), it was

proposed that in a collectivist culture like Turkey, the normative nature of the employment

relationship would generate expectations for loyalty to the organization, and the perceived

costs of violating these expectations would be reflected in increased continuance commitment.

In particular, it was expected that the endorsement of generalized norms for loyalty to one’s

organization and informal recruitment would lead to higher levels of continuance commitment.

The investigation involved two phases. In Study I, in-depth interviews were conducted with

Turkish employees to develop emic items for the scales of interest. In Study II, the proposed

model was tested using structural equations modeling. The results not only confirmed the

cross-cultural generalizability of the antecedents and consequences of affective commitment,

but also indicated that loyalty norms and ingroup approval increased continuance

commitment. The influence of norms and the ingroup was stronger for allocentrics.

Furthermore, for allocentrics, continuance commitment was related to more positive job

outcomes. The results underline the importance of normative concerns in understanding

employee attachment in collectivist contexts and also point to a need for a better measurement

of calculative commitment.

r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Organizational commitment; Affective commitment; Continuance commitment; Culture;

Collectivism; Turkey

*Tel.: +90-216-483-9662; fax: +90-216-483-9699.

E-mail address: [email protected] (S.A. Wasti).

0147-1767/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.

PII: S 0 1 4 7 - 1 7 6 7 ( 0 2 ) 0 0 0 3 2 - 9

Dating back to the 1960s, organizational commitment has been measuredand conceptualized in various ways in the North American literature. Onepopular conceptualization of organizational commitment has been to view itas an affective attachment to the organization. Typically measured by theOrganizational Commitment Questionnaire (OCQ), commitment has beendescribed as ‘‘the relative strength of an individual’s identification with andinvolvement in a particular organization’’ (Mowday, Steers, & Porter, 1979,p. 226). Other researchers, on the other hand, have viewed commitment asremaining with the organization due to recognition of costs associated with leaving.Mainly based on Becker’s (1960) side-bet theory, this approach has presentedcommitment to be a less affective and a more calculative concept that is a function ofthe accumulated interests in the organization, such as pensions and seniority.Calculative commitment has been measured by a scale developed by Ritzer and Trice(1969), later modified by Hrebiniak and Alutto (1972). However, the appropriatenessof these measures has been questioned on both conceptual and empirical grounds(Meyer & Allen, 1984).In 1984, Meyer and Allen proposed a bi-dimensional conceptualization of

organizational commitment that drew on these early works, labeling the former viewas affective and the latter as continuance commitment. In 1990, they added a thirdcomponent, namely, normative commitment (Allen & Meyer, 1990). Originallyintroduced by Weiner and Vardi (1980), normative commitment refers to feelings ofobligations to stay with an organization because of the belief that it is the right thingto do. Meyer and Allen (1991) argued that common to these approaches is the viewthat commitment is a psychological state that (a) characterizes the employee’srelationship with the organization, and (b) has implications for the decisions tocontinue or discontinue membership in the organization. Meyer and Allen (1991)further argued that affective, continuance and normative commitment arecomponents, rather than types, of commitment as an employee’s relationship withan organization might reflect varying degrees of all three. Beyond this, however, thenature of these psychological states differs and the three components of commitmentare argued to develop from different antecedents and to have different implicationsfor job-related outcomes other than turnover (see Allen & Meyer, 1996 and Meyer &Allen, 1997, for a review). Affective commitment develops mainly from positivework experiences, such as job satisfaction and organizational fairness, and isassociated with desirable outcomes, such as higher levels of organizationalcitizenship behaviors, and lower levels of withdrawal behaviors like absenteeismand tardiness. Continuance commitment, on the other hand, has been shown to havetwo primary antecedents: Lack of job alternatives and ‘‘side-bets’’, that is, anythingthat increases the cost of quitting, such as investments in the organization in terms oftime, money and effort. Continuance commitment as such, represents a need to staywith the organization and is not related to positive organizational or individualoutcomes. Finally, normative commitment is argued to develop from organizationalcommitment norms that develop pre-entry (through familial and cultural socializa-tion) or post-entry (through organizational socialization) and appears to bepredictive of positive outcomes, albeit not as strongly as affective commitment.

S. A. Wasti / International Journal of Intercultural Relations 26 (2002) 525–550526

Although the value of a normative perspective, particularly in collectivist cultures,is increasingly recognized (e.g., Meyer & Allen, 1997), the organizationalcommitment literature has been dominated by the former two perspectives (Shore,Tetrick, Shore, & Barksdale, 1999). Empirical research in particular has stronglyvalidated the measures and the value of affective commitment as evidenced in themeta-analysis by Mathieu and Zajac (1990): in 103 out of the 174 samples included intheir study, commitment had been measured by the OCQ. The literature oncontinuance commitment, however, has been less fruitful. Indeed, the results ofCohen and Lowenberg’s (1990) meta-analysis of 50 studies that empirically testBecker’s side-bet model provide little support for his theory of commitment. Yet,several scholars have argued for the usefulness of the side-bet approach, noting thatsome of the limitations of the past research, both in terms of measuring continuancecommitment and the strategies used to test the side-bet model, may be responsiblefor the lack of empirical support (Meyer & Allen, 1984; Shore et al., 1999; Wallace,1997).The purpose of the present investigation is to test an integrated model of affective

and continuance commitment in Turkey. Turkey, as a predominantly collectivist,developing country, represents an interesting as well as an informative contrast tothe individualistic, industrialized context of North America, where most organiza-tional commitment models have been developed and validated. In particular, thecontention of this paper is that whereas affective commitment is an etic (universal)component of commitment, developing primarily from positive work experiencesand predicting desirable organizational and individual outcomes, the nature ofcontinuance commitment varies across cultures. In other words, it is argued that theculturally differing meaning of the employment relationship has implications forwhat constitute side-bets in different contexts. Although this notion is implied inBecker’s (1960) original work, the scarcity of cross-cultural research on organiza-tional commitment has led to an incomplete understanding of the influence ofculture. By testing the proposed model of organizational commitment in the Turkishcontext, it will be possible to validate the North American findings regardingaffective commitment as well as to explore some of Becker’s propositions that havebeen overlooked. Hence, it is also hoped that this paper will address some of theconcerns raised in previous research regarding the utility of the side-bet view orcontinuance commitment.

1. Collectivism and organizational commitment

Since Hofstede’s (1980) prominent investigation, the influence of cultural valueson organizational behavior has been of growing interest to many scholars. Of thefour dimensions identified in Hofstede’s (1980) original work, individualism–collectivism has received the most extensive theoretical development and empiricalvalidation (Triandis, 1995) both as a cultural variable and an individual differencevariable (e.g., Triandis, Chan, Bhawuk, Iwao, & Sinha, 1995; Wagner, 1995). Indeed,this dimension has been found to be an important explanatory variable, accounting

S. A. Wasti / International Journal of Intercultural Relations 26 (2002) 525–550 527

for many differences observed in terms of work-related attitudes, human resourcesmanagement practices and organizational group behavior (e.g., Earley, 1993; Hui,Yee, & Eastman, 1995; Ramamoorthy & Carroll, 1998).The essential difference between individualism and collectivism is with respect to

the construal of the self (Markus & Kitayama, 1991; Triandis, 1995). Inindividualism, the self is viewed as independent, meaning that individuals seek tomaintain their independence from others by giving priority to their own thoughts,needs and feelings and by focusing on their unique attributes. In collectivism, thedefinition of the self is interdependent resulting in an emphasis on attending to othersand adjustment of behavior to maintain a harmonious interdependence with them.These differences are posited to have important implications for social behavior(Markus & Kitayama, 1991), including the employment relationship and the natureof organizational commitment (Randall, 1993; Redding, Norman, & Schlander,1994). In particular, for individualists, the priority of individual goals over groupgoals implies that the employee’s commitment to the organization exists only to theextent that the individual feels that it is to his or her advantage (Allen, Miller, &Nath, 1988). ‘‘Labor is a factor of production in a rational schema. It can beterminated without moral transgression’’ (Redding et al., 1994, p. 653). Thus, theemployer–employee relationship is viewed as a business contract based on thecalculation of mutual advantages. In collectivist cultures, however, the relationshipcontains moral overtones, as with the traditional mutual obligations in an extendedfamily. Protection and loyalty are reciprocal, and change of employment is sociallydisapproved of, as is firing for poor performance (Hofstede, 1991). While there is awide range of types of employer–employee relations within collectivist andindividualist societies, there appears to be a tendency for overlaps of thecollectivist/moral and the individualist/calculative components (Hofstede, 1991;Redding et al., 1994).Despite growing evidence on the existence and importance of these differences,

much of the accumulated literature on psychology and related disciplines such asorganizational behavior is based on the ‘‘Western’’ view of the individual as anindependent, autonomous entity who behaves as a consequence of internal attributessuch as abilities, traits, motives and values (Markus & Kitayama, 1991; Triandis,1994). Arguably, this bias is evident also in the organizational commitment literatureand has resulted in an extensive investigation of affective and calculative approachesto commitment, reflecting a concern for individual needs, wants, preferences andcost-benefit calculations. Normative considerations in the employment relationship,which are particularly relevant for collectivist contexts, have remained relativelyunexplored. In the remaining of this section, Becker’s side-bet view and the conceptof continuance commitment will be revisited and several hypotheses will be advancedto develop a model of organizational commitment in a collectivist context.In line with the individualistic notion of the employment relationship as a business

contract that can be terminated without moral infringement, previous research in theNorth American context has typically treated continuance commitment as reflectinga ‘‘cold calculation of costs and benefits’’ (Jaros, Jermier, Koehler, & Sincich, 1993,p. 954) associated with leaving the organization. As a result, the employee’s


investment in the organization (typically operationalized as age, tenure, hierarchicalposition, or pay) and the availability of alternate job opportunities have beeninvestigated as primary determinants of the continuance or side-bet approach toorganizational commitment (Cohen & Lowenberg, 1990). These conceptualizationsof side-bets are consistent with Becker’s (1960, pp. 36–37) arguments that‘‘impersonal bureaucratic arrangements’’, such as pensions and seniority provisionsas well as ‘‘individual adjustments to social positions’’ that make it difficult to moveto a new job or organization, increase commitment.Yet, Becker (1960, p. 36) has also stated that ‘‘a person sometimes finds that

he has made side bets constraining his present activity because the existenceof generalized cultural expectations provides penalties for those who violatethem. One such expectation operates in the area of work. [y] People feel that aman ought not to change his job too often and that one who does is erraticand untrustworthy.’’ Becker has further argued that side bets are also madethrough the operation of other social processes. Specifically, individuals mightengage in consistent lines of activity out of self-presentation concerns, suchas concerns about living up to other’s expectations or a social image. Thus, Beckerhas in effect proposed that continuance commitment does not develop fromcalculative costs only, but has normative bases as well. As mentioned however, thenotion of normative costs has not been pursued further in the Western literature oncommitment. One very recent notable exception is the scale development study byShore et al. (1999), which does not elaborate on the cultural salience of normativeside-bets but nevertheless takes them into consideration. Incorporating Becker’sframework, it can be argued that in collectivist cultures the prevailing values andnorms that emphasize loyalty to the organization will serve as a side-bet which willbe reflected in higher levels of continuance commitment.The importance of loyalty and trust in the collectivist employment relationship

brings with it a preference for certain managerial mechanisms. One suchmechanism is ingroup or informal recruitment (Adler & Jelinek, 1986). Insteadof screening applicants on the basis of competence, collectivist organizationsoften hire current employees’ friends and relatives. Indeed, in his qualitativeinvestigation on Turkish organizations, Bradburn (1963) observed that alarge proportion of employees got their jobs through relatives and friends.He noted that employers preferred this system because by doing such favors,they increased their employees’ dependency and obligation to the manager.Further, family owned firms had several of the owner’s family on the payrollas the owner felt an obligation to take care of some of the less fortunate membersof his family, even though they may be less capable than someone he might hirefrom outside the family group. Informal recruitment, hence, represents a relationalpsychological employment contract, which creates an obligation to remainloyal (and thus return the favor) on the part of the employee who has beenhired or whose relative has been hired. For the hired employee, the obligation isfurther increased due to a concern to ‘‘maintain the face’’ of the referenceperson whose credibility also is at stake in such a practice. Thus, informalrecruitment can be viewed as a social process where the employee feels bound to


live up to an expectation or an image of a loyal, trustworthy employee or individual.This situation is in line with Becker’s argument that self-presentation concerns mayconstitute a type of side-bet that constrains individual behavior.Based on the above arguments, it was proposed that in a collectivist culture

like Turkey, the normative nature of the employment relationship wouldgenerate expectations for loyalty to the organization, and the perceived costsof violating these expectations would be reflected in increased continuancecommitment. Specifically, it was expected that the endorsement of generalizednorms for loyalty to one’s organization and recruitment through the ingroupwould lead to higher levels of continuance commitment. Furthermore, itwas hypothesized that these relationships would interact with collectivism suchthat the relationships would be stronger for employees who endorse collectivistvalues to a greater extent than employees who do not hold collectivist values asmuch.In addition to these normative bases, in line with the original theory as

well as previous research, it was expected that perceived lack of job alternativesand firm-specific investments, or calculative costs, would also contribute to higherlevels of continuance commitment. However, these side-bets were considered tobe universal and it was not predicted that their salience would be a function ofcultural values. Regarding the outcomes of continuance commitment, it wasexpected that continuance commitment would be negatively related to turnoverintentions. With respect to other outcomes, arguably such a commitment is not aspositive as affective commitment, but it was also expected that this type ofcommitment would not be as unfavorable as continuance commitment conceptua-lized to be instrumental and calculative. Thus, although it was hypothesized that therelationship between continuance commitment and turnover intentions would bestronger for employees who hold collectivist values more than employees who holdcollectivist values to a lesser extent, the tests of the relations to other outcomes suchas citizenship behaviors, work withdrawal and subjective well-being were consideredexploratory.In sum, in the proposed model organizational commitment is conceptualized

as comprising of two components, affective and continuance commitment. Thetwo components are hypothesized to develop from distinct antecedents and tohave differential job-related outcomes. In addition to the hypothesizedantecedents and outcomes of continuance commitment outlined above,consistent with previous American and cross-cultural research (e.g., Allen & Meyer,1996; Ko Price & Mueller, 1997) job satisfaction is proposed as the primaryantecedent of affective commitment. Affective commitment in turn is hypothesizedto predict desirable outcomes, such as higher levels of well-being and or-ganizational citizenship behaviors, and lower levels of turnover intentions and workwithdrawal.The present investigation involved two studies. The purpose of the Study 1 was

twofold: First, to achieve a better understanding of the nature of organizationalcommitment in Turkey, and second, to generate emic (culture-specific) organiza-tional commitment and antecedent items. The generated items were pilot tested and


some of the scales of interest for the present study were developed. In Study 2, theproposed model of organizational commitment was tested using structural equationsmodeling. This model was then compared across individuals who were high or lowon allocentrism (collectivism at the individual level) to further establish the influenceof cultural values on the nature of organizational commitment. These studies and theresults are discussed in the following sections.

2. Study 1: scale development

2.1. Method

2.1.1. Participants and procedure

A fundamental assumption of cross-cultural psychology is that even with perfecttranslation, instruments that are not developed and standardized in that particularculture will yield distorted data as they will not be designed to capture the ‘‘native’spoint of view’’ (Triandis, 1992). Accordingly, the first stage of the study involved thegeneration of emic organizational commitment and antecedent items. The items weregenerated through in-depth interviews conducted with 83 Turkish employees fromvarious organizations (e.g., private, state-owned, family organizations) and differingin terms of hierarchical level, tenure in the organization, gender, age, marital status,education, and rural versus urban background. The interview questions weredeveloped utilizing the framework provided by Triandis’ theory of interpersonalbehavior (Davidson, Jaccard, Triandis, Morales, & Diaz-Guerrero, 1976). Severalitems were generated and classified into five broad categories: generalized norms forloyalty, job satisfaction, obligatory bonding, affective commitment, and continuancecommitment.Next, representative statements for each category were selected and circulated

among the members of a cross-cultural research group comprising of facultymembers and doctorate students who provided further feedback on the categoriza-tion, wording and emic versus etic nature of the items. The final set of items was thenadministered in a pilot survey. The pilot survey also included the affective andcontinuance organizational commitment scales by Meyer, Allen, and Smith (1993),the INDCOL scale (Singelis, Triandis, Bhawuk, & Gelfand, 1995) and other scalesfor validation (described in greater detail below), such as job satisfaction (JobDescription Index [JDI]; Smith, Kendall, & Hulin, 1969) and job withdrawal(Hanisch & Hulin, 1990, 1991). The sites for the pilot study were five governmentalagencies and one public university in Ankara, Turkey. The sample consisted of 351respondents in various clerical, supervisory or professional jobs. Male respondentsaccounted for 46.7% of the sample. The mean levels of age were 33.8 for male and38.8 for female respondents. In terms of education, 86.9% of the male and 85% ofthe female respondents held at least a university (bachelor’s) degree. The averagetenure with the current organization was 8.8 and 13.1 years for the male and femalerespondents, respectively.


2.2. Results

2.2.1. Development of the organizational commitment scale

To develop the organizational commitment scale, initially an exploratory factoranalysis, using maximum likelihood estimation followed by varimax rotation, wascarried out using both the emic and the Meyer et al. (1993) organizationalcommitment items. Only three of the Meyer et al. (1993) continuance commitmentitems were considered to represent this construct in line with the controversy over thedimensionality of this scale. Several researchers have shown that Meyer et al.’s(1993) continuance commitment scale can be broken down into two distinctdimensions of (a) personal costs associated with leaving and (b) awareness of lack ofalternatives (Dunham, Grube, & Castaneda, 1994; Hackett, Bycio, & Hausdorf,1994; McGee & Ford, 1987). The fact that the two subdimensions correlate veryhighly has generated questions as to the meaningfulness of the breakdown (Dunhamet al., 1994). However, Ko et al. (1997) argued that items that reflect personal costsassociated with leaving the organization are congruent with Becker’s (1960) theoryand that the lack of alternatives items should be treated as an antecedent tocontinuance commitment. In light of these arguments continuance commitment wastreated as high personal costs associated with quitting.The exploratory factor analysis supported the two-factor solution and after

analyzing for internal consistency the organizational commitment scale wasfinalized. The final scale consisted of emic items as well as several items by Meyeret al. (1993), which had good psychometric properties and were in fact rather similarto items generated through the interview responses (e.g., ‘‘I feel like part of thefamily in this organization’’). Similarly, some emic items are reminiscent of items inother commitment scales of American origin; however, Turkish employeesmentioned these items frequently and as a result they were considered to be emic.Next, using LISREL 8 (J .oreskog & S .orbom, 1993), confirmatory factor analysis

was carried out to determine whether a two-factor model of affective andcontinuance commitment provided an acceptable fit to the data. Specifically, atwo-factor model was compared with a one-factor model where all items loaded onone factor. Table 1 shows that the improvement in the fit statistics of the two-factormodel over the one-factor model is highly significant (Dw2 (1, N ¼ 335)=177.89,po0:001). Table 2 presents the standardized factor loadings for the final set of itemsthat constituted the scales. Although one continuance commitment item of Meyer

Table 1

Goodness-of-fit indices for confirmatory factor analyses for organizational commitment scales

Model n w2 df w2=df GFI AGFI NNFI SRMSR

One-factor 335 410.05 90 4.56 0.83 0.77 0.65 0.10

Two-factor 335 232.16 89 2.61 0.91 0.89 0.84 0.07

Note: GFI=goodness-of-fit index; AGFI=adjusted goodness-of-fit index; NNFI=non-normed fit index;

SRMSR=standardized root mean square residual.


et al. (1993) had a relatively lower loading, it was kept to maintain the breadth of thescale. On the whole, all emic items have satisfactory loadings and the modificationindices do not suggest major improvement if any parameters are freed.

2.2.2. Analysis of the antecedent items

The next step involved the exploratory factor analysis of the antecedent itemsgenerated from the interviews to develop scales for the proposed side-bets. Interviewitems regarding job satisfaction and perceived lack of alternatives had not beenincluded in the survey. Instead, job satisfaction was measured with the JDI (Smithet al., 1969) as the measurement equivalence of this scale in Turkish was establishedpreviously (Wasti, Bergman, Glomb, & Drasgow, 2000). Likewise, it was decidedthat perceived lack of alternatives would be measured using the relevant subscale ofMeyer et al.’s continuance commitment scale. The remaining antecedent items werefactor analyzed using maximum likelihood estimation followed by varimax rotation.A five-factor solution was obtained as presented in Table 3. The first factor consistedof items referring to generalized loyalty norms and the second factor reflectedinvestments in the current organization. The third factor was named informalrecruitment as it was composed of two items that referred to the obligation thatarises due to informal recruitment. The fourth factor consisted of items that referredto the organization’s investment in the employee, and this factor was labeledorganizational collectivism. The fifth factor reflected the influence of the employee’singroup (family in this case) on his or her decision to stay with an organization.

Table 2

Standardized factor loadings for the two-factor model of organizational commitment

Item

ACS1. I really feel as if this organization’s problems are my own 0.66

ACS2. I do not feel a strong sense of belonging to my organization {R} 0.66

ACS3. I do not feel ‘‘emotionally attached’’ to this organization {R} 0.66

ACS4. I do not feel like ‘‘part of the family’’ at my organization {R} 0.44

ACS5. This organization has a great deal of personal meaning for me 0.72

EAC1. I feel a sense of ‘‘ownership’’ for this organization 0.59

EAC2. I feel it is prestigious to be a part of this organization 0.59

EAC3. I identify with the goals of this organization 0.55

CCS1. It would be very hard for me to leave my organization right now, even if I

wanted to

0.57

CCS2. Too much in my life would be disrupted if I decided I wanted to leave my

organization now

0.55

CCS3. If I had not put so much of myself into this organization, I might consider

working elsewhere

0.28

ECC1. It would be difficult for me to get used to a new organization 0.60

ECC2. I would not want to start from scratch at another organization 0.42

ECC3. The longer I stay with this organization, the harder it is to leave 0.46

ECC4. There is no guarantee that other places will be better, at least I know this

place

0.40

Note: ACS and CCS items are from Meyer et al. (1993) and EAC and ECC items are Emic.


Table 3

Exploratory factor analysis of emic antecedent items: five-factor solution

Item

It is not a good qualification for an

employee to have changed jobs

frequently

0.879 0.088 0.047 �0.104 0.087

It is not good to change jobs

frequently

0.707 0.206 �0.070 �0.049 0.181

If I were an employer I would not

hire someone who has changed

organizations too often

0.635 0.071 0.060 �0.080 0.206

If an employee voluntarily changes

firms, I would think that his/her

behavior is opportunistic or

individualistic

0.340 0.083 0.136 �0.060 0.001

I have spent many years in this

organization learning everything

about its operations

0.081 0.759 �0.141 0.101 �0.052

I am used to this place 0.153 0.577 0.071 �0.086 0.098

I have a certain degree of power as

a senior member

0.106 0.408 0.131 �0.244 0.023

My retirement is due soon 0.054 0.346 0.089 0.231 0.102

I prefer to stay with this

organization as I know the people

inside out

0.300 0.327 0.292 �0.015 0.229

My boss is relying on me 0.060 0.309 0.002 �0.192 0.047

I feel an obligation to stay with my

current organization because I owe

to the person who was my reference

to find this job

0.075 0.032 0.763 �0.049 0.248

I made an implicit contract with

management at the time of my job

acceptance that I would be a long-

term employee

0.063 0.023 0.638 0.006 �0.058

This organization has provided me

neither with financial nor with

emotional support {R}

�0.034 0.046 �0.023 0.649 �0.126

I do not think this organization is

investing in me {R}

0.003 �0.074 0.067 0.547 0.024

My employer looks after the

employees almost like a father

0.113 0.134 0.262 �0.450 0.126

My family thinks it would be

difficult for me to get used to

another organization

0.057 0.014 0.010 0.032 0.541

My family thinks this organization

is the best for me

0.103 0.033 0.055 �0.089 0.528

This organization provides benefits

to my family as well

0.115 0.092 0.058 �0.079 0.306


Although these results support the proposed antecedents of affective andcontinuance commitment to a large extent, two of the factors obtained suggestthat there may be some additional antecedents worth investigating. One of theseantecedents is the factor labeled organizational collectivism. In fact, it is notsurprising that this factor emerged as this type of management philosophy is morecommon in collectivist contexts than in the West (Nam, 1995). Since such managerialpractices are likely to create a surrogate family, it was hypothesized thatorganizational collectivism would increase affective commitment. The second factorconsists of the items that reflect ingroup (family) influence or approval. Based on theinterviews, it appears that informal recruitment practices create an obligation notonly towards the employer but also the family, who has used its personal networks toobtain such a favor. Even in instances where the family was not very instrumental inthe hiring process, many interviewees nevertheless expressed accountability to theingroup and made it clear that their employment decision would be in line with theiringroups’ expectations and approval. Since according to Becker’s theory the notionof feeling a need to live up to others’ expectations contributes to a cost-basedcommitment, it was proposed that ingroup expectations would also serve as anothertype of normative side-bet in the collectivist context.Based on these observations, the proposed model was slightly revised and is

presented in Fig. 1. According to the final version of the model, affective

Affectivecommitment

Continuancecommitment

Satisfactionwith life

Organizationalcitizenshipbehaviors

Turnoverintentions

Workwithdrawal

Satisfactionwith work

Norms forloyalty

Lack ofalternatives

Ingroupinfluence

Informalrecruitment

Investments

Organizationalcollectivism

Fig. 1. Antecedents and consequences of affective and continuance commitment.


commitment is proposed to develop as a function of job satisfaction andorganizational collectivism. Ingroup influence, on the other hand, is hypothesizedto be an antecedent of continuance commitment. As with other culturally salientantecedents of continuance commitment, it was argued that the relationship betweeningroup influence and continuance commitment would be moderated by culture suchthat this relationship would be stronger for allocentric individuals.

3. Study II: testing the two-component model of organizational commitment

3.1. Method

3.1.1. Participants and procedure

This phase involved the test of the two-component model of organizationalcommitment using data from a sample of private sector employees (N ¼ 916). Thesample comprised of 404 females (45.3%) and 487 males (54.7%). The modal agecategory was 25–29 years. The sample was highly educated: 74% of the respondentshad furthered their education beyond high school. Most of the respondents wereoffice workers (32.5%) and only 8% of the sample comprised of blue-collar workers.Eleven percent of the respondents were technicians, 19.5% were supervisors, 15%were professionals, and finally 14% were managers. The modal tenure category was1–5 years (45%).

3.1.2. Measures

Data was collected by a questionnaire in Turkish. Turkish translations of many ofthe scales originally in English had been used successfully in previous research (Wastiet al., 2000). The other scales were translated into Turkish by the author and back-translated by two bilingual academicians. The items that had discrepancies werecorrected and back-translated once again. The scales in the actual survey wereordered so that the dependent variables of interest did not precede all theindependent variables (Podsakoff & Organ, 1986). Unless indicated, all items hada response scale ranging from 1 (strongly disagree) to 7 (strongly agree). Higherscores indicated higher levels of endorsement of the construct.Organizational commitment was measured by the affective and continuance

commitment scales developed in Study 1. As mentioned before, the continuancecommitment scale consisted only of the Meyer et al. (1993) items measuring highpersonal sacrifice associated with leaving. The other three Meyer et al. (1993) itemsplus one new item constituted the scale for perceived lack of alternatives.Satisfaction with work was measured by JDI scale (Smith et al., 1969) as revised

by Roznowski (1989). In the JDI, participants respond Yes, ?, or No to a series ofadjectives describing the characteristics of their work. Organizational collectivismwas measured by a revised version of the scale developed by Robert (1998). Thisscale assesses organizational culture along the individualist–collectivist dimensions.Respondents are presented with a number of statements describing a culturalattribute of an organization (e.g., ‘‘Management and supervisors are protective of


and generous to loyal workers’’). For the purposes of the current investigation, onlythe organizational collectivism scale was employed.The respondents’ endorsement of generalized norms for organizational loyalty

was assessed using three items that were generated in Study 1 and one additionalitem. Likewise, rather than using tenure, or age as a proxy variable, investments inthe organization were measured by the items generated in Study 1 plus one new item.The scale for the influence of the ingroup (family) was measured by six items thatreflect the family’s preference with respect to an individual’s employment decision aswell as the family’s values with respect to organizational loyalty. Three of these itemswere developed in Study 1 and three more were written to ensure adequate reliability.To measure informal recruitment two Yes-No questions, and two multiple-choice

questions were designed. The first Yes-No question asked whether the employee hadan ingroup member (such as family member, relative, spouse, close friend or friendof family) in the organization before he or she started working for the organization.The second Yes-No question asked whether the employee had an ingroup membercurrently employed in the organization. The first multiple-choice question asked howthe employee had gathered information about the organization before joining it. Therespondent was provided with the following options: Through a relative; through afriend/acquaintance; through a newspaper advertisement; and other. The respondentwas instructed to specify what constituted the ‘‘other’’ option. Finally, the secondmultiple-choice question asked how the employee was selected into the organization.The following response options were provided: Through a relative’s reference;through a friend/acquaintance’s reference; through an exam; through an interview;and other. Again, the respondent was instructed to specify what constituted the‘‘other’’ option. The responses to the ‘‘other’’ category were coded as formal orinformal recruitment. The responses to the four questions were summed up such thathigher scores reflected higher levels of informal or ingroup recruitment.Regarding outcomes of commitment, turnover intentions were measured using

three items of the Job Withdrawal scale by Hanisch and Hulin (1990, 1991). Probst(1998) has shown that the original scale is composed of two three-item subscales,namely turnover intentions and ease of quitting. The scale has a 5-point multiplechoice response scale where higher scores reflect stronger turnover intentions. Workwithdrawal was measured with the scales developed by Hanisch and Hulin (1990,1991). The 12-item scale assesses the frequency of behaviors that avoid tasksassociated with one’s work (e.g., taking long breaks, missing meetings) on an 8-pointscale ranging from ‘‘Never’’ to ‘‘More than once a week’’.Since the Work Withdrawal scale has personal industry or productivity related

items, the organizational citizenship behaviors of interest were altruism towardscolleagues and promotion of the organization. To measure these extra-role behaviorsthe self-report scales by Moorman and Blakely (1995) were employed. These itemswere mixed with the Work Withdrawal items, and consequently, the respondentsindicated the frequency of extra-role behaviors (e.g., defending the organizationwhen other employees criticize it, voluntarily helping new employees settle into thejob) on an 8-point scale ranging from ‘‘Never’’ to ‘‘More than once a week’’. Finally,subjective well-being was measured using the 5-item Satisfaction with Life Scale


(SWLS) by Diener and his colleagues (Diener, 1984; Diener, Emmons, Larsen, &Griffin, 1985). The SWLS asks respondents to indicate the extent to which they aresatisfied with their life in general.The respondents were also administered the INDCOL (Singelis et al., 1995) as a

measure of allocentrism and idiocentrism. INDCOL is composed of four 8-itemsubscales, namely Horizontal Individualism, Vertical Individualism, HorizontalCollectivism, and Vertical Collectivism. Triandis and Gelfand (1998) have shownthat the individualism and collectivism constructs can be further delineated ashorizontal (emphasizing equality) or vertical (emphasizing hierarchy). It should benoted that, based on the results of Study 1, some INDCOL items were rewritten inorder to make them easier to understand. For example, the original item ‘‘I like myprivacy’’ was rewritten to mean ‘‘I enjoy being by myself’’ because there is no exacttranslation for the word ‘‘privacy’’. Also, instead of the original item ‘‘When Isucceed it is because of my abilities’’ whose direct translation was not easy to grasp, anew item ‘‘I prefer to make my own decisions rather than consult others’’ wasincluded.Due to the fact that all the variables were measured at the same time and from the

same person, concern over the effects of common method variance was warranted.To test whether such a problem existed, four distractor items on leadership wereintermingled with the INDCOL items. The results indicated that the variables ofinterest did not correlate with these unrelated items.

3.1.3. Data analysis

To test the model presented in Fig. 1, structural equation modeling asimplemented by LISREL 8 (J .oreskog & S .orbom, 1993) was used. As recommendedby Anderson and Gerbing (1988), a two-step structural equation modeling approachwas employed. The first step involved fitting a measurement model that is, assessingthe fit of the observed variables to the latent constructs. As the second step, thestructural model, which establishes the relationships between latent constructs, wasestimated. Both the measurement models and the structural models were estimatedusing maximum likelihood estimation. Covariance matrices were used as the inputfor the LISREL analyses. Three multi-item composites were created for each latentconstruct. The use of multi-item composites reduces the total number of items to amanageable level and provides indicators with higher reliability than that of singleitems (MacCallum, Roznowski, & Necowitz, 1992). The items among the multi-itemcomposites were distributed in an attempt to balance both classical test theory itemstatistics and item content; this procedure maximizes the extent to which theindicators of each construct share variance (Fitzgerald, Drasgow, Hulin, Gelfand, &Magley, 1997).In order to obtain reasonable and stable parameter estimates the analyses

comparing individual differences in the cultural values of collectivism (i.e.,allocentrism) were conducted using a median split of the sample, which ensuredhaving around 300 people with complete data in each group. All analyses wereconducted using a median split on horizontal allocentrism (HC). The choice of thehorizontal dimension of allocentrism primarily was made out of statistical reasons.


Although horizontal allocentrism and horizontal idiocentrism were unrelated to eachother and positively related to vertical allocentrism and vertical idiocentrism,respectively, the latter two scales (vertical allocentrism and vertical idiocentrism)were positively correlated. Thus, it appeared that with the current sample, there wasmore evidence for the construct validity of the horizontal scales. Hence, in theremainder of the study, allocentrism refers to horizontal collectivism measured at theindividual level.Similar to the analyses carried out with the integrated model, the first step

involved the estimation of measurement models for the two groups separately. Afterestablishing that the overall fit of the measurement models was satisfactory, thestructural model was fitted to the two samples independently. Then the model wasestimated simultaneously with an invariance constraint placed on the paths from theexogenous variables to the endogenous variables (gammas, i.e., the paths fromantecedents of commitment to the two commitment constructs) and the paths fromthe endogenous variables to the other endogenous variables (betas; i.e., the pathsfrom the two commitment constructs to the consequences). The final step involvedthe fitting of a partially constrained model where paths reflecting the theorizeddifferences were freely estimated across the two groups. The relative fit of thepartially constrained model was then compared to the fit of the fully constrainedmodel. If the fit of the partially constrained model was found to be significantlybetter than that of the fully constrained model, it was concluded that cultural valuesmoderate the proposed relationships. These comparisons were primarily made byassessing the change in w2 relative to the degrees of freedom. The following sectionsdescribe the results of these steps.

3.2. Results: testing the integrated model of organizational commitment

Prior to testing the model, the reliability of the scales was assessed and a few itemswere eliminated from further analyses. Next, confirmatory factor analyses wereconducted with each scale and satisfactory fit indices were obtained (not reported forbrevity). Table 4 presents the means, standard deviations, coefficient alphas, andcorrelations for the final version of the scales.

3.2.1. Measurement model

The goodness-of-fit indices obtained for the model in Fig. 1 are presented inTable 5. The w2=df ratio is 2.04, which is very good for this sample size (N ¼ 648after listwise deletion). The Goodness-of-fit index (GFI) is 0.91, the ad-justed goodness-of-fit index (AGFI) is 0.89, the non-normed fit index (NNFI) is0.92, the standardized root mean square residual (SRMSR) is 0.047 and the rootmean square error of approximation (RMSEA) is 0.040. To further examine themodel’s fit the residuals and modification indices were inspected; however, nosystematic patterns of misfit were identified. The t-statistics obtained for all thefactor loadings were statistically significant (po0:01). Overall, the results indicate asatisfactory fit.


Table4

Descriptivestatisticsandintercorrelationsofvariables

No.

of

item

s

MeanSD

Alpha

12

34

56

78

910

11

12

13

14

15

16

17

1.Affective

commitment

835.96

9.490.83

2.Continuance

commitment

728.70

8.050.77

0.43nn

3.Satisfaction

withwork

915.74

7.650.84

0.50nn

0.15nn

4.Investm

ents

524.83

5.230.71

0.35nn

0.38nn

0.28nn

5.Lack

of

alternatives

39.64

3.420.53

0.06

0.38nn�0.03

0.07n

6.Norm

sforloyalty

418.16

4.150.51

0.33nn

0.35nn

0.17nn

0.26nn

0.16nn

7.Inform

al

recruitment

42.69

2.470.76

�0.04

0.08n

�0.04

0.01

0.10nn

0.06

8.Ingroupinfluence

414.00

5.850.80

0.30nn

0.56nn

0.12nn

0.21nn

0.36nn

0.28nn

0.15nn

9.Organizational

collectivism

624.62

7.790.79

0.54nn

0.23nn

0.39nn

0.17nn

0.13nn

0.25nn�0.04

0.24nn

10.Work

withdrawal12

14.41

10.560.74

�0.30nn�0.18nn�0.26nn�0.11nn�0.02

�0.18nn�0.01

�0.13nn�0.19nn

11.Turnover

intentions

37.76

2.660.77

�0.53nn�0.35nn�0.45nn�0.17nn�0.15nn�0.28nn

0.01

�0.32nn�0.40nn

0.33nn

12.Organizational

citizenship

10

41.83

14.870.85

0.45nn

0.23nn

0.26nn

0.20nn

0.01

0.16nn

0.04

0.19nn

0.29nn�0.21nn�0.22nn

13.Satisfaction

withlife

520.83

5.700.76

0.28nn

0.14nn

0.11nn

0.14nn

0.04

0.11nn�0.12nn

0.06

0.16nn

0.02

�0.23nn

0.16nn

14.Horizontal

Individualism

839.97

6.360.63

�0.14nn�0.02

�0.07n

�0.04

�0.04

0.01

�0.03

�0.08n

�0.07n

0.13nn

0.14nn�0.02

0.10nn

15.Horizontal

Collectivism

740.42

4.670.68

0.29nn

0.20nn

0.12nn

0.21nn

0.02

0.24nn

0.06

0.17nn

0.18nn�0.25nn�0.10nn

0.37nn

0.12nn

0.05

16.Vertical

Collectivism

729.28

7.070.67

0.15nn

0.39nn

0.02

0.16nn

0.25nn

0.30nn

0.17nn

0.39nn

0.14nn�0.20nn�0.08*

0.15nn

0.04

�0.11nn

0.38nn

17.Vertical

Individualism

733.55

6.270.64

0.16nn

0.15nn

0.09*

0.20nn

0.09nn

0.28nn

0.06

0.11nn

0.08n

�0.03

�0.04

0.20nn

0.07n

0.23nn

0.20nn

0.20nn

npo0:05:

nn

po0:01(two-tailed

test).


3.2.2. Structural model

As shown in Table 5, the fit indices for the structural model included a w2=df ratio2.35, a GFI of 0.89, an AGFI of 0.87, a NNFI of 0.89, a SRMSR of 0.063 and anRMSEA of 0.046, indicating an acceptable fit. Fig. 2 illustrates LISREL estimates ofthe structural model coefficients. The paths from satisfaction with work andorganizational collectivism to affective commitment are positive and significant (0.37and 0.56, respectively). Of the proposed antecedents to continuance commitmentconceptualized as high personal sacrifice, only informal recruitment did not reachsignificance (�0.05). Otherwise, as Becker (1960) proposed, it appears that thegreater the endorsement of generalized norms of loyalty, the higher is the level ofcontinuance commitment (0.20). Further, the path coefficient from ingroup influenceto continuance commitment indicates that higher levels of family preference orapproval for continued employment in the current organization is related to higherlevels of perceived costs associated with quitting (0.37). Fig. 2 illustrates that the‘‘etic’’ antecedents are also significantly related to continuance commitment. Thepath coefficient of 0.29 shows that higher levels of investments in the organizationare associated with increased perceptions of costs. Likewise, perceived lack ofalternatives is positively associated with continuance commitment (0.34).Except for turnover intentions, no firm predictions were made with respect to the

outcomes of continuance commitment. As predicted, both components ofcommitment are significantly negatively related to turnover intentions. The pathfrom affective commitment to desirability of quitting is �0.64 and from continuancecommitment the coefficient is �0.15. As far as other consequences of commitmentare concerned, as predicted, affective commitment is significantly related tosatisfaction with life in general (0.44); however, continuance commitment isunrelated to life satisfaction (�0.05). Thus, employees who express a desire tomaintain membership to their organization also report higher levels of well-being.On the other hand, employees who feel a need to stay due to high costs associatedwith leaving do not to report positive or negative judgments regarding their lives.

Table 5

Goodness-of-fit indices for measurement and structural models

w2 df w2=df GFI AGFI NNFI SRMSR RMSEA

Measurement models

Full sample 1270.14 624 2.04 0.91 0.89 0.92 0.047 0.040

High allocentrics 1069.08 624 1.71 0.86 0.83 0.89 0.057 0.047

Low allocentrics 959.63 624 1.54 0.87 0.84 0.91 0.057 0.041

Structural models

Full sample 1562.24 666 2.35 0.89 0.87 0.89 0.063 0.046

Submodel I: High allocentrics 1233.11 666 1.85 0.84 0.81 0.87 0.069 0.051

Submodel II: Low allocentrics 1136.27 666 1.71 0.85 0.82 0.88 0.074 0.047

Note: GFI=goodness-of-fit index; AGFI=adjusted goodness-of-fit index; NNFI=non-normed fit index;

SRMSR=standardized root mean square residual; RMSEA=root mean square error of approximation.


Affective commitment is significantly negatively related to work withdrawal(�0.36), but the path from continuance commitment did not reach significance(�0.06). Likewise, although affective commitment is significantly positively relatedto extra-role or organizational citizenship behaviors (0.47), continuance commitmentis unrelated (0.06). Thus, for the most part, it appears that although continuancecommitment, operationalized as high personal sacrifice, is influenced by norms andthe ingroup similar to normative commitment, it nevertheless is a less productiveform of commitment.

3.3. Results: fitting the model across high versus low allocentrism groups

3.3.1. Measurement models for high versus low allocentrics

The comparison across the groups began by fitting separate measurement modelsto the two data sets. As presented in Table 5, the goodness-of-fit indices obtained forthe high allocentrism group are a w2=df ratio of 1.71, a GFI of 0.86, an AGFI of0.83, a NNFI of 0.89, a SRMSR of 0.057 and an RMSEA of 0.047. In addition,there appeared to be no discernable pattern of residuals, and the modification indices


Affectivecommitment

Continuancecommitment: Highpersonal sacrifice



Informalrecruitment

Norms

Ingroupinfluence

Investments

Workwithdrawal


Workwithdrawal

Organizationalcitizenship

Turnoverintentions


Turnoverintentions

Lack ofalternatives

.37* (.05)

.56* (.05)

-.05 (.04)

.20* (.06)

.37* (.06)

.29* (.05)

.34* (.07)

.44* (.05)

-.36* (.05)

.47* (.05)

-.64* (.05)

-.05 (.05)

-.06 (.05)

.06 (.04)-.15*(.04)

Fig. 2. Structural model coefficients (and their standard errors). Note: *po0:05 (one-tailed).


did not imply a problematic specification. The goodness-of-fit indices for themeasurement model for the low allocentrism group also appeared to be satisfactory.Table 5 shows that a w2=df ratio of 1.54, a GFI of 0.87, an AGFI of 0.84, a NNFI of0.91, a SRMSR of 0.057 and an RMSEA of 0.041 was obtained for this group. Inlike manner, examination of standardized residuals and the modification indicesconfirmed that the fit was acceptable. The factor loadings for each set of threeindicators obtained for the two samples were all significant (po0:01).

3.3.2. Structural models for high versus low allocentrics

Firstly, the structural model was estimated separately for the two groups. Aspresented in Table 5, satisfactory fits were obtained for each sample. For the highallocentrism sample, the w2=df ratio was 1.85, the GFI 0.8, the AGFI 0.81, the NNFI0.87, the SRMSR 0.069 and the RMSEA was 0.051. The corresponding values werew2=df ratio of 1.71, a GFI of 0.85, an AGFI of 0.82, a NNFI of 0.88, a SRMSR of0.074 and an RMSEA of 0.047 for the low allocentrism sample. Although SRMSR ishigher than desirable, the standardized residuals and the modification indices were ofacceptable magnitudes and there were no discernible patterns that suggestedsignificant improvement if the model was respecified. Fig. 3 provides the commonmetric solution for the individually estimated models.The next step involved estimating the model simultaneously across the two groups

with an equality constraint placed on the structural coefficients. Table 6 shows thatthe fit of the constrained model was acceptable (w2=df ratio of 1.78, a GFI of 0.84, aNNFI of 0.88, SRMSR of 0.074). Since the freely estimated and the fully constrainedmodels are nested, their difference in w2 was compared to their difference in degreesof freedom and a ratio of 75.85/41=1.85 was obtained. Thus, there does not seem tobe a substantial decrement in fit when the paths are constrained to be equal acrossthe groups.The first partially invariant model involved testing for influence of cultural values

on the antecedents proposed to be culturally salient. Accordingly, the paths fromgeneralized norms for loyalty and ingroup influence were freely estimated across thetwo groups. It should be noted that informal recruitment was not included in theanalysis as this path did not reach significance in either sample. Table 6 shows thatthe decrease in w2 was 6.07 for 2 df, which is significant at po0:05: The paths fromnorms for loyalty and ingroup influence to normative commitment are 0.29 and 0.39for individuals who endorse allocentrism to a greater extent; the correspondingcoefficients are 0.08 and 0.34 for individuals that are low on allocentrism. Thedifference seems primarily in the way norms predict continuance commitment: Forindividuals that are low on allocentrism, there is no a significant relation betweenthese two variables suggesting that the existence of generalized norms for loyaltydoes not increase the perceived costs of quitting. For highly allocentric individualson the other hand, norms for loyalty significantly increase the perceived costs ofquitting.Next, the paths from continuance commitment to job-related outcomes were freely

estimated across groups. Since the path from continuance commitment toorganizational citizenship behaviors was zero for both samples, these paths were


not freed. Table 6 shows that the improvement in the model’s fit was modest butsignificant (w2 decreased by 9 for 3 df). The path coefficient from continuancecommitment to intention to leave is significantly negative (�0.23) for individuals


Affectivecommitment

Continuancecommitment

Normsfor loyalty

Informalrecruitment


Ingroupinfluence

Investments

Lack ofalternatives

.33* (.06).40* (.07)

.52* (.08)

.57* (.07)

.19* (.10)

.44* (.07)-.01 (.09)-.08 (.05)

.11 (.10).24* (.08)

.31* (.07)

.30* (.07)

.60* (.13)

.24* (.08)

.59* (.08)

.31* (.06)

-.30* (.08)-.37* (.06)

.49* (.08).44* (.06)-.70* (.09)

-.60* (.06)

-.20* (.07)

.06 (.07)

.02 (.06)-.11* (.07)

.04 (.06).01 (.07)-.08 (.06)

-.24* (.06)



Turnoverintentions


Workwithdrawal

Turnoverintentions


Workwithdrawal

Fig. 3. Structural model coefficients (and their standard errors) for the individually estimated models.

Note: *po0:05 (one-tailed). Regular font represents path coefficients for high allocentrics (n ¼ 329); bold

font represents path coefficients for low allocentrics (n ¼ 319).

Table 6

Goodness-of-fit indices for the fully constrained and partially constrained models: allocentrism split

Model w2 df w2=df Dw2=Ddf GFI NNFI SRMSR RMSEA

Separately estimated 2369.38 1332 1.78 0.85 0.88 0.074 0.035

Fully constrained 2445.23 1373 1.78 0.84 0.88 0.074 0.035

Partially constrained model I 2439.16 1371 1.78 3.04n 0.84 0.88 0.074 0.035

Partially constrained model II 2436.23 1370 1.78 3n 0.84 0.88 0.075 0.035

npo0:05:GFI=goodness-of-fit index; NNFI=non-normed fit index; SRMSR=standardized root mean square

residual; RMSEA=root mean square error of approximation.


high on allocentrism but is not significant for individuals that do not hold collectivistvalues strongly (�0.08). Although no firm prediction was made, the results also showthat continuance commitment is significantly negatively related to work withdrawal(�0.12) for individuals high on allocentrism but unrelated to work withdrawal forthe low allocentrism sample. Interestingly, continuance commitment is significantlynegatively related to satisfaction with life for the latter group (�0.14) suggesting thata need to stay with the organization creates feelings of frustration to some extent.This path is not significant for the high allocentrism sample.

4. Discussion

The major goals of the present study were to explore the influence of collecti-vist values on the nature of organizational commitment as well as to evaluatethe generalizability of the findings in the Western literature with respect to thedimensionality, development and consequences of commitment. Most of theresearch on organizational commitment has focused on affective commitment andthis type of commitment has been validated across various samples (e.g., Ko et al.,1997). Indeed, this study also supported that affective commitment was a function ofpositive work experiences, associated with desirable work and personal outcomes,and that these relationships were not different across individuals with differingendorsement of collectivist cultural values. Therefore, in the remainder of thisdiscussion, the cross-cultural generalizability of the side-bet approach to organiza-tional commitment will be evaluated.The present investigation was primarily triggered by the observation that the

dominance of North American research in the area of organizational commitment,as in other topics of organizational behavior and psychology, has resulted in anindividualistic interpretation of Becker’s framework. With the exchange relationshipas the accepted or prevailing form of an employment contract, valued side-bets, forthe most part, have been conceptualized as material investments in the organizationand foregoing such investments has been identified as the major cost associated withquitting. This is not to say that material investments are important only inindividualistic contexts. Quite the contrary, the argument of the present investigationwas that these costs are indeed universal. What may be cultural, however, are thenorm-based costs associated with the employment relationship. The cross-culturalliterature suggests that such costs may be more relevant for collectivist cultureswhere the employment relationship is more personal and long-term oriented,resulting in the formation of an obligatory bonding. It is interesting to note thatBecker (1960) explicitly stated that for a complete understanding of a person’scommitments, one needs an analysis of the system of values or valuables in the worldthe individual lives in. In Becker’s words, it is necessary to discern what kind ofthings are conventionally wanted and what losses feared in a society in order tounderstand the employee–employer bond in that context. This paper contrasted thevalue systems of individualist and collectivist cultures to achieve a betterunderstanding of organizational commitment in a collectivist culture like Turkey.


The results indicated that for Turkish employees continuance commitmentdeveloped not only as a result of investments in the organization and perceived lackof alternatives but also from generalized norms for loyalty and the approval of theemployee’s ingroup. The influence of collectivist values on the nature of continuancecommitment was more strongly confirmed when further analysis revealed that forindividuals low on collectivism there was no relation between generalized norms forloyalty and continuance commitment. For individuals high on collectivism, however,loyalty norms significantly increased continuance commitment. However, althoughinformal recruitment was hypothesized to create a pressure to live up to an image ofa loyal employee and consequently increase continuance commitment, the results didnot support this expectation. One possible explanation is that the measurement ofinformal recruitment was not optimal. Another possibility is that introducing theingroup approval variable might have accounted for the variance associated with theinformal recruitment variable. During the interviews, it was observed that in manycases informal recruitment channels involved the family’s connections and not justthe employee’s. For instance, one interviewee stated: ‘‘The board member who wasmy reference person to get this job is a very precious family friend. Knowing myfamily and me, he recruited me because he felt I could help fulfil his mission for thisorganization. If I were to quit and let him down, our family relations would beruined’’. Another interviewee said, ‘‘The Vice President is my uncle. My parents feelmore at ease and less worried when they know there is someone to watch over me inmy workplace’’. Thus, it appears that in several cases the obligations aroundinformal recruitment are reflected in the family’s preferences. Arguably, the familymight be more important in getting the first job for a young person. Considering thatthe 72% of the sample is between the ages of 20–34 and that 70% have 5 or less yearsof tenure, the importance of the family may have been amplified. However, ascurrently around 75% of the Turkish population is below the age 35, sampling olderage groups is rather difficult. Nevertheless, the results support the notion that undercertain contexts, employee attachment will develop in such a way that affect orinstrumental gains will have little to do with it.With respect to outcomes, for individuals high on collectivism continuance

commitment was related to lower levels of turnover intentions and work withdrawal(e.g., neglecting assigned duties, tardiness). For individuals low on collectivism,continuance commitment did not predict any positive outcomes. It is interesting tonote that continuance commitment, which reflects a need to stay, seems to be a moreproductive form of attachment for individuals high on collectivism.

5. Limitations and future research directions

The present investigation addresses a dearth in the organizational commitmentliterature by attempting to portray the influence of cultural values on the attachmentprocess. The study employed sound cross-cultural methodology and the proposedmodel was tested rigorously, by taking into account both antecedents and outcomesof two dimensions of commitment. Nevertheless, the cross-sectional, single-country,


survey research design has inherent weaknesses and some of the findings are not asstrong as desired. Therefore, the results should be considered exploratory but theissues they raise suggest fruitful and important future research avenues. These issuesare discussed below.The focus of this paper was the development and implications of continuance

commitment. The current operationalization of continuance commitment representsa need to stay with the organization due to high personal sacrifice associated withquitting. When measured as such, in line with Becker’s original work, it can beargued that affective, normative and calculative costs will all serve as antecedents tocontinuance commitment. Indeed, in this paper both calculative and normative costswere found to be predictive of continuance commitment. However, it is believed thatat least three problems emerge with this sort of a conceptualization. Firstly, itbecomes impossible to predict the outcomes of continuance commitment (other thanturnover intentions) unless the nature of the side-bets is taken into account.Secondly, continuance commitment, which takes into account affective andnormative side-bets overlaps with affective and normative commitment as itdevelops from common antecedents. Thirdly, such an operationalization does notreliably represent a calculative attachment to the organization. It is believed thatthese confusions can be curtailed by refining the continuance commitment scale. Inparticular, to the extent that normative and affective concerns are captured with theother two components, the continuance commitment scale should not consist ofvague items expressing awareness of unspecified costs but only items that reflectcalculative or instrumental attachment. This will allow the testing of a three-component model of commitment, where each commitment component developsfrom unique antecedents and has different implications for outcomes.Future research effort should also focus on the development of the norma-

tive perspective to attachment, both theoretically and empirically. Althoughnormative commitment was not included in the present study, it was shown thatnormative concerns create feelings of a need (or an obligation) to stay with thecurrent organization. These findings suggest that at least in collectivist contexts, thedevelopment of organizational commitment will not be captured fully unlessnormative concerns are also taken into account. The development of apsychometrically sound normative commitment scale will allow a better testing ofhypotheses raised in this study. Another interesting and urgent research avenue is theconsequences of normative commitment. So far, the little empirical researchavailable has suggested that normative commitment is related to desirable jobbehaviors, albeit not as strongly as affective commitment. However, Meyer andAllen (1997) have also noted that feelings of obligation might occasionally bringabout feelings of resentment, and employees who experience high levels of normativecommitment might carry out desirable work behaviors grudgingly. A usefuldistinction in terms of understanding the consequences of normative commitmentmay involve viewing normative commitment as comprising of two components:Personal norms and social determinants. In fact, such a distinction would allowviewing the commitment process to be parallel to Triandis’ theory of interpersonalbehavior (1980). According to this theory, an individual’s behavioral intention is a


function of: (1) his/her affect toward performing the act; (2) his/her beliefs about theconsequences of performing that behavior and the evaluation of those consequences;(3) social factors, that is, the perceived appropriateness of a particular behavior formembers of specific reference groups (normative beliefs) and occupants of specificpositions in the social structure; and (4) the subject’s personal normative beliefsabout what he or she should or ought to do with regard to the behavior of interest(personal norms). Drawing on the current results, one could argue that normativecommitment that develops as a consequence of internalized personal norms wouldlead to desirable outcomes but normative commitment developing out feelings ofobligation to others, that is, social factors, would be more likely to createresentment.To conclude, although much research has accumulated in the area of

organizational commitment, there is still a need for further conceptual refinementand empirical validation. It is believed that more cross-cultural work will alsocontribute towards a clarification of some of the processes underlying employeecommitment.

Acknowledgements

This paper is based on the author’s dissertation. The author would like toacknowledge the exemplary supervision of her thesis advisor Harry C. Triandis aswell as the invaluable guidance of Fritz Drasgow. Further acknowledgment is due tothe Office of International Studies and the Institute of Labor and IndustrialRelations, University of Illinois at Urbana-Champaign for their financial support.

References

Adler, N. J., & Jelinek, M. (1986). Is ‘organizational culture’ culture bound? Human Resources

Management, 25, 73–90.

Allen, N. J., & Meyer, J. P. (1990). The measurement and antecedents of affective, continuance and

normative commitment to the organization. Journal of Occupational Psychology, 63, 1–18.

Allen, N. J., & Meyer, J. P. (1996). Affective, continuance, and normative commitment to the

organization: An examination of construct validity. Journal of Vocational Behavior, 49, 252–276.

Allen, D. B., Miller, E. D., & Nath, R. (1988). North America. In R. Nath (Ed.), Comparative management

(pp. 23–54). Cambridge, MA: Ballinger.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and

recommended two-step approach. Psychological Bulletin, 103, 411–423.

Becker, H. S. (1960). Notes on the concept of commitment. American Journal of Sociology, 66, 32–42.

Bradburn, N. M. (1963). Interpersonal relations within formal organizations in Turkey. Journal of Social

Issues, 31, 61–67.

Cohen, A., & Lowenberg, G. (1990). A re-examination of the side-bet theory as applied to organization

commitment: A meta-analysis. Human Relations, 43, 1015–1050.

Davidson, A. R., Jaccard, J. J., Triandis, H. C., Morales, M. L., & Diaz-Guerrero, R. (1976). Cross-

cultural model testing: Toward a solution of the etic-emic dilemma. International Journal of

Psychology, 11, 1–13.

Diener, E. (1984). Subjective well-being. Psychological Bulletin, 95, 542–575.


Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of

Personality Assessment, 49, 409–420.

Dunham, R. B., Grube, J. A., & Castaneda, M. B. (1994). Organizational commitment: The utility of an

integrative definition. Journal of Applied Psychology, 79, 370–380.

Earley, P. C. (1993). East meets west meets mideast: Further explorations of collectivistic and

individualistic work groups. Academy of Management Journal, 36, 319–348.

Fitzgerald, L. F., Drasgow, F., Hulin, C. L., Gelfand, M. J., & Magley, V. J. (1997). The antecedents and

consequences of sexual harassment in organizations: A test of an integrated model. Journal of Applied

Psychology, 82, 578–589.

Hackett, R. D., Bycio, P., & Hausdorf, P. A. (1994). Further assessments of Meyer and Allen’s (1991)

three-component model of organizational commitment. Journal of Applied Psychology, 79, 15–23.

Hanisch, K. A., & Hulin, C. L. (1990). Job attitudes and organizational withdrawal: An examination of

retirement and other voluntary behaviors. Journal of Vocational Behavior, 37, 60–78.

Hanisch, K. A., & Hulin, C. L. (1991). General attitudes and organizational withdrawal: An evaluation of

a causal model. Journal of Vocational Behavior, 39, 110–128.

Hofstede, G. (1980). Culture’s consequences. Beverly Hills: Sage.

Hofstede, G. (1991). Cultures and organizations: Software of the mind. London: McGraw-Hill.

Hrebiniak, L. G., & Alutto, J. A. (1972). Personal and role-related factors in the development of

organizational commitment. Administrative Science Quarterly, 17, 555–573.

Hui, C. H., Yee, C., & Eastman, K. L. (1995). The relationship between individualism-collectivism and job

satisfaction. Applied Psychology: An International Review, 44, 276–282.

Jaros, S. J., Jermier, J. M., Koehler, J. W., & Sincich, T. (1993). Effects of continuance, affective, and

moral commitment on the withdrawal process: An evaluation of structural equation models. Academy

of Management Journal, 36, 951–995.

J .oreskog, K. G., & S .orbom, D. (1993). LISREL 8: User’s reference guide. Chicago: Scientific International

Software.

Ko, J. W., Price, J. L., & Mueller, C. W. (1997). Assessment of Meyer and Allen’s three-component model

of organizational commitment in South Korea. Journal of Applied Psychology, 82, 961–973.

MacCallum, R. C., Roznowski, M., & Necowitz, L. B. (1992). Model modifications in covariance

structure analysis: The problem of capitalization on chance. Psychological Bulletin, 111, 490–504.

Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and

motivation. Psychological Review, 98, 224–253.

Mathieu, J. E., & Zajac, D. M. (1990). A review and meta-analysis of the antecedents, correlates, and

consequences of organizational commitment. Psychological Bulletin, 108, 171–194.

McGee, G. W., & Ford, R. C. (1987). Two (or more?) dimensions of organizational commitment:

Reexamination of the affective and continuance commitment scales. Journal of Applied Psychology, 72,

638–642.

Meyer, J. P., & Allen, N. J. (1984). Testing the side-bet theory of organizational commitment: Some

methodological considerations. Journal of Applied Psychology, 69, 372–378.

Meyer, J. P., & Allen, N. J. (1991). A three-component conceptualization of organizational commitment.

Human Resources Management Review, 1, 61–89.

Meyer, J. P., & Allen, N. J. (1997). Commitment in the workplace: Theory, research and application.

Thousand Oaks: Sage.

Meyer, J. P., Allen, N. J., & Smith, C. A. (1993). Commitment to organizations and occupations: Some

methodological considerations. Journal of Applied Psychology, 78, 538–551.

Moorman, R. H., & Blakely, G. L. (1995). Individualism-collectivism as an individual difference predictor

of organizational citizenship behavior. Journal of Organizational Behavior, 16, 127–142.

Mowday, R. T., Steers, R. M., & Porter, L. W. (1979). The measurement of organizational commitment.

Journal of Vocational Behavior, 14, 224–247.

Nam, S. H. (1995). Culture, control and commitment in international joint ventures. International Journal

of Human Resources Management, 6, 553–567.

Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and prospects.

Journal of Management, 12, 531–544.


Probst, T. (1998). Antecedents and consequences of job insecurity: An integrated model. Unpublished

Doctoral Dissertation, University of Illinois, Urbana.

Ramamoorthy, N., & Carroll, S. (1998). Individualism/collectivism orientations and reactions toward

alternative human resource management practices. Human Relations, 51, 571–588.

Randall, D. (1993). Cross-cultural Research on organizational commitment: A review and application of

Hofstede’s value survey module. Journal of Business Research, 26, 91–110.

Redding, S. G., Norman, A., & Schlander, A. (1994). The nature of individual attachment to the

organization: A review of East Asian variations. In: H. C. Triandis, M. Dunnette, & L. Hough (Eds.),

Handbook of industrial and organizational psychology, Vol. 4 (2nd ed.) (pp. 557–607). Palo Alto, CA:

Consulting Psychologists Press.

Ritzer, G., & Trice, H. M. (1969). An empirical study of Howard Becker’s side-bet theory. Social Forces,

47, 475–479.

Robert, C. A. (1998). The development and validation of the horizontal and vertical individualism and

collectivism organizational culture scale. Unpublished Doctoral Dissertation, University of Illinois,

Urbana-Champaign.

Roznowski, M. (1989). An examination of the measurement properties of the job description index with

experimental items. Journal of Applied Psychology, 74, 805–814.

Shore, L. M., Tetrick, L. E., Shore, T. H., & Barksdale, K. (1999). Construct validity of measures of

Becker’s side-bet theory. Poster presented at the annual meeting of the Society for Industrial and

Organizational Psychology, Atlanta.

Singelis, T. M., Triandis, H. C., Bhawuk, D., & Gelfand, M. J. (1995). Horizontal and vertical dimensions

of individualism and collectivism: A theoretical and measurement refinement. Cross-Cultural Research:

The Journal of Comparative Social Science, 29, 240–275.

Smith, P. C., Kendall, L., & Hulin, C. L. (1969). The measurement of satisfaction in work and retirement.

Chicago: Rand McNally.

Triandis, H. C. (1992). Cross-cultural research in social psychology. In D. Grandberg, & G. Sarup (Eds.),

Social judgment and intergroup relations: Essays in honor of Muzafer Sherif. New York: Springer.

Triandis, H. C. (1994). Cross-cultural industrial and organizational psychology. In: H. C. Triandis,

M. Dunnette, & L. Hough (Eds.), Handbook of industrial and organizational psychology, Vol. 4 (2nd

ed.) (pp. 103–172). Palo Alto, CA: Consulting Psychologists Press.

Triandis, H. C. (1995). Individualism and collectivism. Boulder, Co: Westview Press.

Triandis, H. C., Chan, D. K.-S., Bhawuk, D. P. S., Iwao, S., & Sinha, J. B. P. (1995). Multimethod probes

of allocentrism and idiocentrism. International Journal of Psychology, 30, 461–480.

Triandis, H. C., & Gelfand, M. (1998). Converging measurement of horizontal and vertical individualism

and collectivism. Journal of Personality and Social Psychology, 74, 118–128.

Wagner III, J. A. (1995). Studies of individualism-collectivism: Effects on cooperation in groups. Academy

of Management Journal, 38, 152–172.

Wallace, J. E. (1997). Becker’s side-bet theory of commitment revisited: Is it time for a moratorium, or a

resurrection? Human Relations, 50, 727–749.

Wasti, S. A., Bergman, M. E., Glomb, T. M., & Drasgow, F. (2000). Test of the cross-cultural

generalizability of a model of sexual harassment. Journal of Applied Psychology, 85, 766–778.

Weiner, Y., & Vardi, Y. (1980). Relationships between job, organization, and career commitments and

work outcomes: An integrative approach. Organizational Behavior and Human Performance, 26, 81–96.


Affective and continuance commitment to the organization: Evaluation of measures and analysis of...

Documents

Transcript of Affective and continuance commitment to the organization: Evaluation of measures and analysis of...