RESEARCH AND PRACTICE
IN HUMAN RESOURCE MANAGEMENT

Highlight, copy & paste to cite:

Kuruvilla, S., & Sverke, M., (1993). Two Dimensions of Union Commitment Based on the Theory of Reasoned Action: Cross – Cultural Comparisons, Research and Practice in Human Resource Management, 1(1), 1-16.

Two Dimensions of Union Commitment Based on the Theory of Reasoned Action: Cross – Cultural Comparisons

Sarosh Kuruvilla & Magnus Sverke

Abstract

Following Friedman and Harvey’s (1986) call for additional theory development of the union commitment construct, this study applies the theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975) to the study of union commitment. The theory is tested using questionnaire data from diverse samples of 1,914 professional union members from Sweden and 1,088 blue collar union members from the United States of America. Confirmatory factor analysis results of LISREL 7 indicate support for the two dimensions of union commitment posited by the theory: Union Attitudes and Opinions, and Pro-Union Behavioural Intentions. The construct validity of the dimensions is demonstrated through correlational data. The similarity of results across the diverse samples provides evidence of cross-cultural generalisability of the dimensions, and suggests support for the application of the theory of reasoned action to union commitment research.

Introduction

The interest in the structure of the union commitment construct began with a study conducted by Gordon, Philbot, Burt, Thompson, and Spiller (1980). Based on earlier work on organisational commitment (Dubin, Champoux, & Porter, 1975; Mowday, Porter, & Steers, 1982), Gordon et al (1980) defined union commitment as the extent to which an individual (a) has a strong desire to remain a member of the union; (b) is willing to exert high levels of effort on behalf of the union, and; (c) believes in and accepts the goals of the union. In their study of United States white-collar workers, Gordon et al (1980) developed a 484-item measure of union commitment and identified four dimensions of the concept: Union Loyalty, Willingness to Work for the Union, Responsibility to the Union, and Belief in Unionism.

While the four-dimensional structure introduced by Gordon et al (1980) has been supported in numerous replicating studies (eg Gordon, Beauvais, & Ladd, 1984; Ladd, Gordon, Beauvais, & Morgan, 1982; Tetrick, Thacker, & Fields, 1989; Thacker, Fields, & Tetrick, 1989) other studies (eg Fullagar, 1986; Kiandermans, 1986; Mellor, 1990) have uncovered different factor structures. For example, based on a re-analysis of the Gordon et al (1980) data, Friedman and Harvey (1986) suggested that union commitment is best represented by two dimensions, namely Union Attitudes and Opinions, and Pro-Union Behavioural Intentions.

Noting that their two-factor solution did not “easily mesh” with the theoretical view of union commitment suggested by Gordon et al (1980), Friedman and Harvey (1986) called for additional theory development of the union commitment construct as well as for further psychometric modification of the Gordon et al (1980) questionnaire (cf Gallagher & Strauss, 1991). While several researchers have answered Friedman and Harvey’s (1986) call for additional refinement of the Gordon et al (1980) questionnaire (eg Fullagar, 1986; Klandermans, 1986; Ladd et al, 1982; Tetrick et al, 1989; Thacker et al, 1989), their call for additional theory development of union commitment is, as yet, not completely answered.

Given the current controversy over the number of dimensions of the concept, additional theory-based approaches to the study of union commitment are justified for several reasons. First, as Walkey and McCormick (1985) note, the best choice of number of dimensions of any construct should be obtained from some theoretically derived source. Second, once a theoretical basis for union commitment is established, a theory-based approach permits the researcher to test the ability of a hypothesised structure to fit the data, unlike much of the previous research that has relied mainly on exploratory factor analysis. Finally, it provides a stable means for meaningful comparisons of factor structures across different samples and cultures. Currently, with a few exceptions (eg Tetrick, Note 5), researchers generally uncover differing dimensions of union commitment in diverse settings by means of exploratory factor analysis, but are unable to draw firm conclusions as to why the factors of union commitment differ across samples and national/cultural boundaries.

A crucial issue in the number-of-dimensions debate concerns the fact that the factors of union commitment deal with different dimensions of time. The Gordon et al (1980) definition of union commitment as well as its original source — the predominant definition of organisational commitment (Dubin et al, 1975; et al 1982) — both include attitude components as well as desire-to-remain-a-member and willingness-to-exert-extra-efforts components. As Mathieu and Zajac (1990) note, the issue of whether these “desire components” are distinguishable from other variables like, for example, behavioural intentions, warrants further examination. It seems, therefore, desirable to distinguish the attitudinal process nature of commitment from its outcome dimensions (Guest & Dewe, 1991; Sverke, Note 3).

The Theory of Reasoned Action

The theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975) constitutes an interesting theoretical guideline for obtaining a more parsimonious factorial representation of union commitment. Since union commitment has been conceptualised as an attitude (eg Fullagar & Barling, 1987), an application of the Fishbein and Ajzen (1975) attitudinal theory to making decisions on the dimensionality of union commitment seems justified. A fundamental underpinning of the theory is the assumption that human beings are rational; before people act, they make systematic use of information available to them and consider the consequences of their actions, and then they decide either to engage or not to engage in social action (Ajzen & Fishbein, 1980, p. 5).

The theory of reasoned action is based on the distinction drawn between the three components of any attitudinal construct, cognition, affect, and conation (Allport, 1935). In Fishbein and Ajzen’s (1975) framework, these components translate into beliefs, attitudes, and intentions. Briefly, according to the theory, beliefs lead to the formation of attitudes; attitudes — along with subjective norms, ie, the tendency to conform to the opinions of salient others — lead to the formation of intentions to perform various behaviours; intentions, then, are the immediate determinants of action. Consequently, attitudes are general predispositions; they do not predispose individuals to perform specific behaviours, but rather lead to sets of behavioral intentions (Ajzen & Fishbein, 1980, p.5).

The distinction between beliefs and attitudes is, however, more theoretical than practical. More specifically, Fishbein and Ajzen (1975) argue that measures of beliefs will often yield results similar to measures of attitudes since both concepts are usually assessed by direct self reports. As Bern (1965; 1972) suggests, measures of beliefs and attitudes are usually self-descriptive verbal responses, which are influenced by both external and internal stimuli. One source of these stimuli for attitudinal responses is the person’s opinions and behaviour, as well as the context in which these occur. Thus, in effect, the beliefs and attitudes that an outsider would attribute to the individual are functionally equivalent in that both sets of statements are inferences from the same evidence. Therefore, as long as the assessment instruments use self report measures, in practical terms there are only two components to an attitudinal sphere: beliefs/attitudes and behavioural intentions.

The separation of attitudes from intentions in commitment research has received both theoretical and empirical support. For example, it has been suggested that the definition of commitment should leave out the intentions and, instead, focus on the psychological attachment to the organisation (O’Reilly & Chatman, 1986) or the identity with the organisation (Guest & Dewe, 1991). Further, although Friedman and Harvey (1986) in their re-analysis of the Gordon et al (1980) data did not explicitly examine the theory of reasoned action, they concluded that they “would expect the Fishbein model to predict the observed separation of behavioural intentions from the affective and attitudinal items” (Friedman & Harvey, 1986, p. 376). Other studies (Kiandermans, 1986; Mellor, 1990) have provided further support for Friedman and Harvey’s (1986) two-dimensional representation of union commitment. The theory of reasoned action has also been tested in several areas of the social sciences such as union voting behaviour (Montgomery, 1989) and organisational commitment (Horn & Hulin, 1981; Horn, Katerberg & Hulin, 1979).

Purpose and Hypotheses

Utilising the theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975) to model the dimensions of union commitment, the objective of the present study was to examine the dimensionality of union commitment . The study also aimed at establishing the construct validity and generalisability of the dimensions suggested by the theory.

An application of the theory of reasoned action to the study of union commitment suggests some testable hypotheses relating to the dimensionality of union commitment and the construct validity of these dimensions. First, the theory implies that there are two dimensions of union commitment, one being attitudinal and one consisting of behavioural intentions. Second, the theory implies that these two dimensions, although related, are in fact distinct constructs. Third, since in the Fishbein and Ajzen (1975) framework beliefs are precursors to attitudes, general beliefs about unions should evidence a closer relationship with the attitudinal dimension of union commitment than with the intentional dimension. Similarly, measures of attitudes (job and union satisfaction) should be more closely related to the attitudinal commitment measure relative to the dimension of behavioural intentions. Fourth, since, according to the theory of reasoned action, intentions are more strongly related to behaviour than are attitudes, the behavioural intentions should evidence a stronger correlation with union directed behaviour than with the attitudinal dimension of commitment.

We use the confirmatory factor analysis methods of LISREL 7 to test whether the two dimensions of union commitment implied by the application of the theory of reasoned action can be identified, The validity and generalisability of the dimensions are then examined using conventional procedures, but in diverse samples of professional union members from Sweden, and blue-collar postal workers from the U.S. The diversity of the samples also distinguishes this study from previous approaches in that it provides an opportunity to test a theory, based approach of union commitment across national cultures and occupational categories. If the pattern of results is similar across both samples, then this will constitute evidence of generalisability of the dimensions of union commitment to diverse populations and cultural contexts.

Method

Subjects and Procedure

Participants in this study were Swedish professionals (affiliated with the Swedish Union Federation for Professional Employees, SACO), and letter carriers employed by the U.S. Postal Service (affiliated with the National Association of Letter Carriers, NALC). The subjects were randomly selected from lists of union members provided by SACO and NALC, and mail questionnaires were distributed to 2,900 and 2,400 individuals in the Swedish and the U.S. samples respectively. For the Swedish sample, translation of the questionnaire was required and the method of translation — back translation was used. A cover letter explaining the purpose of the study was sent along with the questionnaire to the subjects in both samples. In the Swedish sample, the cover letter from the President of the union requested members to cooperate with the survey as it was primarily to be used for “international and comparative research purposes”; in the U.S. sample, the letter noted that the survey was required for the union to get an idea of member opinions regarding grievances and workplace issues. In both cases, no follow-up mailings or contacts were attempted.

The response rates were 66 % for the Swedish sample and 45 % for the U.S. sample resulting in 1,914 and 1,088 usable questionnaires respectively for the two samples. The professionals in the Swedish sample represented 26 different occupations such as university professors, doctors, lawyers, clergymen, military officers, economists, and psychologists. The mean age of the Swedish sample was 42 years (SD=14.21) and the average union tenure was 11.8 years (SD=8.63). Of the Swedish respondents, 59 % were men. In the U.S. sample, the mean age was 37.52 years (SD=12.24) and the average union tenure was 12.2 years (SD=8.8). Males comprised 87 % of the U.S. sample.

Measures

Largely similar questionnaires were used in both samples although there were minor variations in some of the items in order to take into account cultural and occupational differences between the samples, and also some differences between the questionnaires due to limitations set by the unions. Besides demographics (age, sex, and union tenure), data were collected about the respondents’ beliefs about unionism in general, their job and union satisfaction, their union participation, and their union commitment. Unless else specified, the variables were measured with Likert-type scales ranging from 1 (strongly disagree) to 5 (strongly agree). Reversed items were recoded prior to analysis and indexes were constructed using mean values of the items in the scale.

General beliefs about unions. In the Swedish sample, general beliefs about unions were measured by nine statements related to the extent to which the respondents agreed that unions are instrumental in obtaining wage and benefit increases, increasing job security for their members, are harmful to the economy, and are responsible for inflation and breakdown in labour-management relationships. The mean was 3.16 (SD=.46) and the reliability estimate (Cronbach’s alpha) was .79 for the Swedish sample. In the U.S. sample, this variable was measured by a single item concerning the extent to which the respondents value unionism in general, a measure that is closely related to beliefs about unions.

Union satisfaction. In both samples, union satisfaction was measured by multi-item Likert-type scales. In the Swedish sample, the scale contained eight statements about the extent to which individuals agreed that they were satisfied with the union’s performance on various job-related dimensions such as pay, benefits, internal member-union relations, efforts to improve the quality of work life, and the ability of union officials in carrying out their responsibilities (alpha=.83). The scale used in the U.S. sample primarily reflected the respondents’ satisfaction with the way union stewards carried out various union tasks (alpha=.78). Mean values were 2.97 (SD=.56) and 3.26 (SD=.66) for the Swedish and the U.S. sample respectively.

Job satisfaction. Shorter versions of the Minnesota Satisfaction Questionnaire were used to measure job satisfaction. The scale measured the respondents’ degree of satisfaction with a number of factors related to the job and the responses were made on a 5 point scale ranging from 1 (very dissatisfied) to 5 (very satisfied). 13 items were used in the Swedish sample (alpha=.82) while six were used in the U.S. sample (alpha=.79). Mean levels of satisfaction were 3.44 (SD=.61) and 3.40 (SD=.75) for Swedish and U.S. samples respectively.

Union participation. In both samples, a scale comprising eight items was used to assess union participation. Respondents were asked to indicate whether during the last year had been involved in any of the following activities: attending membership meetings, serving as a union steward, serving as an elected branch officer, being a delegate to a union convention, serving on other committees, reading union newsletters regularly, voting in union elections, and campaigning in union elections (items were coded 0=no, 1=yes and summed to create the participation scale). The mean participation level was 1.42 (SD=.28) for the Swedish sample and 1.08 (SD=.49) for the U.S. sample, while the reliabilities (alpha coefficients) were .85 and .70 respectively.

Union commitment. Ten Likert-type items partly derived from Gordon et al (1980) were used to measure union commitment in both samples. Consistent with the theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975) and with the two dimensional nature of union commitment suggested by Friedman and Harvey (1986), the scale was created to reflect two dimensions of union commitment, ie Union Attitudes and Opinions, and Pro-Union Behavioural Intentions. Although there were minor variations in item wordings across samples (some items were worded negatively in one sample and positively in the other), the scale was similar in both surveys. The items are listed in Table 1. Cronbach’s alpha was .78 in the Swedish sample and .80 in the U.S. sample for the attitudinal dimension, and .76 and .70 in the two samples respectively for the intentions dimension. Mean values of the attitudinal dimension were 3.41 (SD=.61) for the Swedish sample and 3.70 (SD=.84) for the U.S. sample, while the means of the intentions dimension were 3.21 (SD=.59) and 3.26 (SD=.66) respectively.

Results

Dimensionality of Union Commitment

The confirmatory factor analysis methods provided by LISREL 7 were used to examine the hypothesised factor structure to the data. LISREL estimates (maximum likelihood) of the factor loadings for the two hypothesised dimensions are reported in Table 1. The factor loadings for both dimensions of union commitment showed a similar pattern across the two samples, with most loadings higher than .50. The two dimensions explained substantial proportions of variance in both samples; Union Attitudes and Opinions accounted for 48% and 50% of the variance in the Swedish and the U.S. samples respectively, while the Pro-Union Behavioural Intentions dimension explained 40% and 42% of the variance in the two samples respectively. LISREL estimates of the correlations (corrected for attenuation) between the two latent factors were .60 for the Swedish sample and .74 for the U.S. sample.

Table 1
LISREL Estimates of Factor Loadings on Union Attitudes and Opinions (UCATT) and Pro-Union Behavioural Intentions (UCINT)
Union Commitment Items Factor Loadings
Swedish Sample U.S. Sample
UCATT UCINT UCATT UCINT
Variance explained 48% 40% 50% 42%
Inter-factor correlation .60 .74
UCATT 1 I talk up this union to my friends as a great organisation to be a member of .71 .82
UCATT 2 I have very little loyalty to the union .70 .44
UCATT 3 Deciding to join the union was a smart move on my part .54 .85
UCATT 4 Quite a lot can be gained by being a member of this union .70 .82
UCATT 5 I trust the union representatives .60 .49
UCINT 1 I am willing to exert considerable effort to make the union successful .81 .87
UCINT 2 It is a member’s duty to support and help other members .35 .59
UCINT 3 It is a member’s duty to ensure that union and members live up to their responsibilities/contracts .44 .40
UCINT 4 I would not do any special work to help this union .71 .73
UCINT 5 If asked, I would run for union office .74 .73

The results of the confirmatory factor analysis are reported in Table 2. For the two-dimensional solution, the adjusted goodness-of-fit indicators (AGFI) were .93 for the Swedish sample and .94 for the U.S. sample; the normed fit index (NFI) was also high in both samples (.93 and .96 for Sweden and the U.S. respectively). On the basis of the criterion that the NFI be greater than or equal to .90 (Bentler & Bonett, 1980), the two-factor model indicated excellent fit for both samples.

Table 2
Confirmatory Factor Analysis Results
Country Model k df chi2k GFI AGFI RMSR NFI PFI
Model Comparisons Sweden USA
df chi2 df chi2
Chi2 difference between Null model and Two factor model 11 5319.15*** 11 4436.16***
Chi2 difference between One and Two dimensional models 1 820.40*** 1 331.53***
Sweden
(n=1675)
Null model 0 45 5751.53 .49 .37 .31
One dimension 1 35 1252.78 .85 .77 .08 .78 .72
Two dimensions 2 34 432.38 .96 .93 .06 .92 .90
USA
(n=1088)
Null model 0 45 4621.48 .38 .24 .39
One dimension 1 35 516.85 .90 .84 .05 .89 .86
Two dimensions 2 34 185.32 .96 .94 .04 .96 .95

*** p<.001

Following conventional procedures, the hypothesised two-factor specification was tested against a null model specification to establish whether the fit of the hypothesised model was better than the fit of the unconstrained null model. The difference between the chi-squares of the two-dimensional model and the null model was significant in both samples (χ2(11) = 5319.15, p < .001 for the Swedish sample and χ2(11) 4436.16, p < .001 for the U.S. sample), suggesting that the two-dimensional model differed substantially from the null model in both samples.

In addition, given the high (disattenuated) correlations between the two dimensions in both samples, the hypothesised two-factor structure was also tested against a single-factor model specification in order to examine whether only one factor underlied the data. In both samples, the AGFI, the NFI, the goodness-of-fit index (GFI, and the parsimonious) fit index (PFI) were substantially higher for the two-dimensional solution than for the unidimensional solution. Further, the differences in chi-squares between the two-dimensional and the unidimensional models were significant for both samples (χ2(1) = 820.40, p < .001 for the Swedish sample and χ2(1) = 331.53, p < .001 for the U.S. sample).

Construct Validity

In order to assess the construct validity of the two dimensions of union commitment, the first-order correlations between the two dimensions and the external attitudinal and behavioural variables were examined. As shown in Table 3, general beliefs about unionism showed a stronger correlation with Union Attitudes and Opinions (UCATT) in both samples (r = .64 for the Swedish sample and r = .63 for the U.S. sample) relative correlation with to the lower Pro-Union Behavioural Intentions (UCINT) in both samples (r = .37 and r = .49 for Sweden and the U.S. respectively). Likewise, union satisfaction evidenced a stronger relationship with the attitudinal dimension (r = .54 for the Swedish sample and r = .73 for the U.S. sample) than with the intentional dimension (r = .18 and r = .37 for Sweden and the U.S., respectively) of union commitment in both samples. In contrast, the attitudinal work-related external variable used in this study — job satisfaction — was only weakly related to the attitudinal dimension of union commitment. While job satisfaction was not significantly related to the Pro-Union Behavioural Intentions dimension in the U.S. sample, it evidenced a negative relationship with the same variable in the Swedish sample. Hotelling’s statistic tests revealed that the correlations with the two union commitment dimensions differed significantly for all external variables.

Table 3
Correlations between Union Commitment Dimensions and Other Variables
Sweden USA
Variable UCATT UCINT ta UCATT UCINT ta
General beliefs about unions .64*** .37*** 15.77*** .63*** .49*** 7.03***
Union satisfaction .54*** .18*** 18.42*** .73*** .37*** 13.85***
Job satisfaction .07* -.07* 4.00*** .16** .03 5.11***
Union participation .32*** .57*** 13.27*** .44*** .54*** 4.63***

*p<.05 **p<.01 ***p<.001

a Hotelling’s t statistic test for differences between correlations of each external variable with Union Attitudes and Opinions (UCATT) and the correlations of each external variable with Pro-Union Behavioural Intentions (UCINT).

In both samples, the behavioural external variable — union participation — was more closely related to Pro-Union Behavioural Intentions (r = .57 for the Swedish sample and r = .54 for the U.S. sample) than to Union Attitudes and Opinions (r = .32 and r = .44 for Sweden and the U.S., respectively). To further examine the relationship between union participation and the two dimensions of union commitment, union participation equations for both samples were estimated with all other variables in the study as independent variables. The result of these equations shows that — after controlling for the effect of Pro-Union Behavioural Intentions — the coefficient for Union Attitudes and Opinions was non-significant in predicting union participation in the U.S. sample, while it was weakly significant (p < .05) in the Swedish sample.

Discussion

Consistent with our first hypothesis, the confirmatory factor analysis results render support for the two dimensions of commitment identified by the application of the theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975) to the study of union commitment. Further, the similarity of results across the samples suggests that the two-dimensional solution has generality over diverse occupations and different national settings. Although Swedish and U.S. unions differ in terms of their goals, structures, and strategies (Abrahamsson, Note 1; Ahlén, Note 2; Kjellberg, 1983), these differences apparently do not translate into differences in the factor structure of union commitment as suggested by the Fishbein and Ajzen (1975) framework.

However, we do not argue that these two factors alone necessarily represent union commitment. While the ten union commitment items used in this study were created to reflect only the two dimensions identified by the theory of reasoned action, alternative theories may suggest different numbers of dimensions. For example, it has been suggested that the definition of commitment ought to focus only on identity with the organisation (Guest & Dewe, 1991), or on value rationality-based and instrumental rationality-based commitment (Sverke, Note 2). Further, an extended set of items like, for instance, the commonly used 30-item version of the Gordon et al (1980) scale, might have created different factor structures.

Although the purpose of our study was to examine the applicability of the theory of reasoned action to union commitment research, a limitation of the study is the inability to test for more than two factors, rendering comparisons with previous studies problematic. The theory of reasoned action presents a theoretical basis for understanding union commitment that differs from previous work on the topic (eg Gordon et al, 1980; Tetrick et al, 1989). Future research must, therefore, further examine the Fishbein and Ajzen (1975) theory in the union commitment context.

Nevertheless, the results of the present study are consistent with the two-dimensional solution suggested by Friedman and Harvey (1986) and also reported by Klandermans (1989) and Mellor (1990). Further, the differing correlations between the union commitment dimensions on one hand, and the external variables on the other, indicate support for the discriminant validity of the two dimensions. As predicted in our second hypothesis, the two dimensions of union commitment, although related, appear to be distinct constructs.

As expected, in both samples the union-related attitudinal variables were more closely related to the attitudinal dimension of union commitment than to the dimension of intentions. Although no examinations of causal relationships were attempted, a priori interpretations of the findings suggest that general beliefs about unions and satisfaction with the unions increase attitudes of commitment to the union, while to a lesser extent influencing the Pro-Union Behavioural Intentions.

In contrast to previous results (eg Gordon et al, 1980), job satisfaction — the attitudinal work-related variable used in this study — was, only weakly related to the attitudinal dimension of union commitment. Consistent with our third hypothesis, job satisfaction was not significantly related to the intentions dimension of commitment in the U.S. sample; unexpectedly, however, it evidenced a negative relationship with the same variable in the Swedish sample. While this relationship was weak and significant only on the 5% level, a very cautious interpretation suggests that Swedish union members may be more willing to work for their union if they are dissatisfied with their jobs, indicating that they are likely to view the union as instrumental in alleviating job dissatisfaction, or suggesting that dissatisfied members may be more likely to turn to active union work. In sum, our third hypothesis developed from the theory of reasoned action receives partial, but not complete, support.

Consistent with our fourth hypothesis, union participation was more strongly related to the Pro-Union Behavioural Intentions dimension than to the attitudinal dimension of union commitment. This was true for both samples. However, the significant correlations between the attitudinal union commitment dimension and union participation noted in both samples contradicts the theory of reasoned action; according to Ajzen and Fishbein (1980), attitudes should affect participation only through behavioural intentions. The correlations between union participation and the attitudinal commitment dimension found in both samples, therefore, indicate that attitudes may influence behaviour through intentions as well as directly.

However, after controlling for the effect of the intentions dimension, the results of the union participation equations indicate that the attitudinal dimension of union commitment was only weakly significant in predicting union participation in the Swedish sample and non-significant in the U.S. sample. This suggests that for Swedish union members, attitudes of union commitment appear to impact participation through intentions as well as directly, although the direct effect is very small. In effect, the results suggest that only the dimension of intentions is really important in predicting behaviour (cf Horn et al, 1979).

The advantage of applying the theory of reasoned action to the study of union commitment — apart from providing guidelines for a decision on the dimensionality, and apart from providing a stable basis for comparisons of dimensions across different samples — is that it highlights the processes underlying the formation of attitudes of commitment. Beliefs are combined with information about the union to develop attitudes of union commitment; attitudes along with subjective norms then influence behavioural intentions, and through intentions, influence behaviour (cf Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975). The theory presents a more dynamic view of union commitment and its formation relative to the more static definitions and models of the concept that have been the basis of much of the research up to now.

However, future research needs to study the processes underlying the formation of union commitment in more detail. Forthcoming studies would benefit from including an extended set of variables and should at least include subjective norms to allow for a test of the full Fishbein and Ajzen (1975) model. Although the present results correspond to the predictions made by the theory and appear generalisable across cultures and occupations, our findings need to be replicated, preferably in a longitudinal setting, before any firm conclusions as to causal directions can be made.

References

Ajzen, I., and Fishbein, M., (1980), Understanding Attitudes and Predicting Social Behavior, Englewood Cliffs: Prentice-Hall.

Ailport, G. W., (1935), Attitudes, in C. Murchison (Ed.), A Handbook of Social Psychology, Clark University Press, Worcester, MA. pp. 798-844.

Bern, D. J., An Experimental Analysis of Self Persuasion. Journal of Experimental Social Psychology, 1, pp.199-218.

Bern, D. J., (1972), Self-perception Theory, in L. Berkowitz (Ed.), Advances in Experimental Social Psychology, Academic Press, New York, pp.1-62.

Bentler, P. M., and Bonett, D. C., (1980), Significance Tests and Goodness-of-fit in the Analysis of Covariance Structures, Psychological Bulletin 88, pp.588-606.

Dubin, R., Champoux, J. E., and Porter, L. W., (1975), Central Life Interests and Organisational Commitment of Blue-collar and Clerical Workers, Administrative Science Quarterly, 20, pp.411-421.

Fishbein, M., and Ajzen, I., Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research, Addison Wesley, Reading, MA.

Friedman, L., and Harvey, R. J., (1986), Factors of Union Commitment: The Case for a Lower Dimensionality, Journal of Applied Psychology, 71, pp.371-376.

Fullagar, C., (1986), A Factor Analytic Study on the Validity of a Union Comniitrnent Scale, Journal of Applied Psychology, 71, pp.129-136.

Fullagar, C., and Barling, J., (1987), Toward a Model of Union Commitment, in D. Lewin and D. Lipsky (Eds.), Advances in Industrial and Labor Relations, 4, JAI Press, Greenwich.

Fullagar, C., and Barling, J., (1989), A Longitudinal Test of the Antecedents and Consequences of Union Loyalty, Journal of Applied Psychology, 74, pp.213-227.

Gallagher, D.G., and Strauss, G., (1991), Union Membership Attitudes and Participation, in G. Strauss, D.G. Gallagher, and J. Fiorito (Eds.), The State of the Unions, Madison, W.I., Industrial Relations Research Association.

Gordon, M.E., Beauvais, L.L. and Ladd, R. T., (1984), The Job Satisfaction and Union Commitment of Unionized Engineers, Industrial and Labor Relations Review, 37, pp.359-370.

Gordon, M. E., Philbot,J. W., Burt, R., Thompson, C. A., and Spiller, W.E., (1980), Commitment to the Union: Development of a Measure and an Examination of its Correlates, Journal of Applied Psychology, 65, pp.479-499.

Guest, D. E., and Dewe, p., (1991), Company or Trade Union: Which Wins Workers Allegiance? A Study of Commitment in UK Electronic Industry, British Journal of Industrial Relations, 29, pp.75-96.

Horn, P. W., and Hulin, C. L., (1981), A Competitive Test of the Prediction of Enlistment by Several Models, Journal of Applied Psychology, 66, pp.23-39.

Horn., P. W., Katerberg, R. J., and Hulin, C. L., (1979), Comparative Examination of Three Approaches to the Prediction of Turnover, Journal of Applied Psychology, 64, pp.280-290.

Kjellberg, A., (1983), Facklig Organisering i Tolv Iänder, Lund: Arkiv, (Ph.D. Thesis).

Kiandermans, B., (1986), Psychology and Trade Union Participation: Joining, Acting, Quitting, Journal of Occupational Psychology, 59 pp.198-204.

Kiandermans, B., (1989) Union Commitment: Replications and Tests in the Dutch Context, Journal of Applied Psychology, 74, pp.869-875.

Ladd, R. T., Gordon, M. E, Beauvais, L. L., and Morgan, R. L., (1982), Union Commitment: Replication and Extension, Journal of Applied Psychology, 67, pp.640-644.

Mathieu, J. E., and Zajac, D. M., (1990), A Review and Meta-analysis of the Antecedents, Correlates, and Consequences of Organizational Commitment, Psychological Bulletin, 1008, pp.171-194.

Mellor, S., (1990), The Relationship between Membership Decline and Union Commitment: A Field Study of Local Unions in Crisis, Journal of Applied Psychology, 75, pp.258-267.

Montgomery, B R., (1989), The Influence of Attitudes and Normative Pressures on Voting Decisions in a Union Cerification Election. Industrial and Labor Relations Review, 42, pp.262-279.

Mowday, R. T., Porter, L. W., and Steers, R. M., (1982), Employee-Organizational Linkages, Academic Press, New York.

O’Reilly, C., and Chatman, J., (1986), Organizational Commitment and Psychological Attachment: The Effects of Compliance, Identification, and Internalization on Prosocial Behavior, Journal of Applied Psychologogy, 71, pp.492-499.

Tetrick, L. E., Thacker, J. W., and Fields, M. W., (1989), Evidence for the Stability of the Four Dimensions of the Commitment to the Union Scale, Journal of Applied Psychology, 74, pp.819-822.

Thacker, J. W.,Fields, M. W., and Tetrick, L., (1989), Evidence of the Four Factors of Union Commitment: An Application of Confirmatory Factor Analysis, Journal of Applied Psychology, 74, pp.228-232.

Walkey, F. H., and McCormick, I. A., (1985), Multiple Replication of Factor Structure: A Logical Solution for a Number of Factors Problem, Multivariate Behavioral Research, 20, Pp.57-67.

Reference Notes

  1. Abrahamsson, B. Union Structural Change. Paper presented at the Symposium on Emerging Union Structures: An International Comparison, Clark University, March 1992.
  2. Ahlén, K. Union legitimacy: Members’ Perceptions of Union Government. Paper presented at the Symposium on Emerging Union Structures: An International Comparison, Clark University, March 1992.
  3. Sverke, M. Developing a Measure of Union Commitment in Sweden: Results of a Pilot Study. Paper presented at the Symposium on Emerging Union Structures: An International Comparison, Clark University, March 1992.
  4. Sverke, M. Toward a New Definition of Union Commitment: The Impact of Instrumental and Value Rationality. Working paper Swedish Center for Working Life.
  5. Tetrick, L. E. Relations Among the Four Dimensions of Union Commitment. Paper presented at the meeting of the Canadian Psychological. Association, 1989.