Union commitment and participation among Hong Kong firefighters: a development of an integrative model

Abstract

This paper examines the factors influencing members' intent to participate in their union, drawing on a study of Hong Kong firefighters. As in Western studies, the main direct predictor of intent to participate is affective union commitment, but instrumentality is a relatively more important antecedent of affective union commitment and participation than in the Western studies. We found some evidence of frustration-aggression effects on intent to participate in ‘militant’ and ‘general’ union activities, although the latter was significant only among individuals who perceive management–employee relations as co-operative and trusting.

Keywords:

• Union commitment
• union participation
• Hong Kong
• industrial relations climate
• firefighters

Notes

Ed Snape (address for correspondence), Department of Management and Marketing, Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, PRC (tel: [phone]; fax: [phone]; e-mail: [email protected]). Andy W. Chan, Department of Management and Marketing, The Hong Kong Polytechnic University, Kowloon, Hong Kong, PRC. Tom Redman, Sheffield University Management School, Sheffield, UK.

 1 Hong Kong appears in Hofstede's (1980, 1983; Westwood, 1992) large power distance-low individualism quadrant, clustered along with Singapore, Malaysia and other Asian countries; in the large power distance-weak uncertainty avoidance quadrant, with Singapore, Malaysia, the Philippines and India; and in a cluster with Singapore and Malaysia as weak uncertainty avoidance, moderate masculinity. It also rates highly on ‘Confucian work dynamismalong with Taiwan, South Korea and Japan.

 2 As with job satisfaction, in addition to this direct effect on affective union commitment, there is a possible indirect effect, via organizational commitment.

 3 Fuller and Hester (1998) find no evidence of moderation for the relationship between affective union commitment and union participation. We anticipate no moderation for this relationship, or for the paths between instrumental union commitment and pro-union attitudes and between job satisfaction and organizational commitment.

 4 Although precise demographic data on the characteristics of the union's membership were not available, discussions with union officers suggested that our sample was broadly representative of the membership. The distribution by rank for the workforce as a whole was 68.53 per cent fireman, 19.65 per cent senior fireman, and 11.82 per cent principal fireman, almost identical with our respondents.

 5 This was done because Hong Kong unions, the Hong Kong Fire Services Department Staffs General Association included, have tended to emphasize the provision of direct benefits to members (e.g. retail discounts).

 6 Before combining items, we satisfied ourselves that each scale was unidimensional by running a series of exploratory factor analyses for the items comprising each scale. In all cases, only a single factor with an eigenvalue greater than one emerged.

 7 In the structural models, we allowed the error terms for the two dimensions of intent to participate in union activities to correlate.

 8 Unlike Fuller and Hester's (1999) meta-analysis, our study is based on a single employer. It is therefore appropriate to interpret our IR climate variable as reflecting perceptual differences.

 9 This path was marginally significant in model A (beta = -0.201;p = 0.046).

10 We repeated the structural analysis using ordinary least squares regression and incorporating the following control variables: age and union tenure in years, marital status (1=married; 0=not married), dummy variables for form 5 graduates and matriculation or degree (other education was the missing category), and dummy variables for senior fireman and principal fireman (fireman was the missing job level category). This analysis provided very similar results to the structural equation modelling, with similar R² values, no differences in the significance/non-significance of regression coefficients and similar-sized coefficients. Thus, the inclusion of control variables had little effect on our findings.

11 In the unconstrained revised model, the estimated standardized total effect coefficient (including direct and indirect paths) of instrumental union commitment on affective union commitment is 0.556, compared to 0.263 for pro-union attitudes.

12 Eight paths were constrained to be equal across the two groups: job satisfaction and instrumental union commitment to affective union commitment and both dimensions of intent to participate, and pro-union attitudes and organizational commitment to affective union commitment.

13 In addition, comparing the path from job satisfaction to intent to participate in militant activity across the two sub-groups yielded a critical ratio of 1.903, just missing statistical significance at the 0.05 level. The relative coefficients were similar to the general union activity case, with that for the less adversarial group being significant and negative (beta = -0.176;p = 0.003) and that for the more adversarial group being non-significant (beta = -0.071;p = 0.202).

14 However, judging from the estimates in the revised model, the standardized total effect of affective union commitment on intent to participate (0.565 for general union activities and 0.484 for militant activities) is larger than that of instrumental commitment (0.404 and 0.387 respectively).

15 Although Fuller and Hester (1998) were considering the effect on union commitment rather than an unmediated effect on participation

16 This was the first time that the full membership of this union had been surveyed in this way, and it was necessary to guarantee anonymity in order to secure participation and achieve a reasonable response rate. It was therefore not possible to match questionnaires to separate reports of actual participation.