The tests of the model indicate that many of the independent variables have different influences on UNCOM and UNSAT.
In the Canadian sample, beliefs about union instrumentality (WORKINST) are positively and significantly related to both UNCOM and UNSAT (p < .
The union information variables appear to have different effects on UNCOM and UNSAT, suggesting a pattern of difference between the two constructs.
The significant relationship noted between these variables and UNCOM is consistent with the notion that UNCOM develops over time, and is more stable than UNSAT (all these variables reflect information obtained in the past).
Co-workers, on the other hand, may not have similar values, but they share experiences in connection with the union; and the positive relationship between COWORKERS and UNSAT may reflect the influence of co-workers' opinions about the union on evaluations of union performance by the respondent.
In the Canadian sample, both measures of job satisfaction, EXTSAT and INTSAT, have strong positive relationships with UNSAT and no significant relationship with UNCOM.
In both countries, the models appear to explain a greater proportion of the variance in UNCOM than in UNSAT, suggesting that the model may be underspecified in the UNSAT equation.
2] for both UNCOM and UNSAT increased by one or two percentage points only.
001) related to UNSAT in Sweden and Canada, INTSAT was negatively related to UNSAT in Sweden.
A test of the hypothesis that each non-intercept (unstandardized) parameter for both dependent variables is the same across samples (that is, in a comparison of the UNCOM equation in Sweden with the UNCOM equation in Canada, and of the UNSAT equation in Sweden with the UNSAT equation in Canada) cannot be rejected for most variables.