Trust and breach of the psychological contract

Administrative Science Quarterly,  Dec, 1996  by Sandra L. Robinson

• E-mail
• Print
• Link

Finally, regressing intentions to remain (T3) on psychological contract breach, number of promotions received, number of pay increases received, trust (T1), intentions to remain (T1), and work experience, psychological contract breach was found to be significant and negatively related to intentions to remain (T3). When unmet expectations was added to the equation, it was significant, but psychological contract breach also remained significant, although its beta weight decreased. Conversely, when trust (T3) was added to the equation, it was significant, but psychological contract breach also remained significant, although its beta weight decreased. The beta weight of psychological contract breach decreased more when trust (T3) was entered into the equation than when unmet expectations was entered into the equation. When both unmet expectations and trust were added to the first equation, both mediators were significant, although trust (T3) explained more variance than unmet expectations, and psychological contract breach was no longer significant. Taken together, this set of regression equations indicates that trust (T3) and unmet expectations (T3) each separately partially mediated the relationship between psychological contract breach and intentions to remain (T3); together the two mediators fully mediated the relationship; and trust (T3), compared with unmet expectations, was a stronger mediator.

Thus the relationship between psychological contract breach and performance was fully mediated by unmet expectations and by trust, although trust explained more variance than unmet expectations. The relationship between psychological contract breach and civic virtue was fully mediated by trust only and not at all mediated by unmet expectations. Finally, the relationship between psychological contract breach and intentions to remain was partially mediated by unmet expectations and by trust, although trust played a stronger mediating role than unmet expectations, and fully mediated by both unmet expectations and trust together. The results show considerable support for the third and fourth hypotheses, which predicted that unmet expectations and trust would mediate the negative relationships between psychological contract breach and employee contributions.

Hypothesis 5 posited that trust (T1) would moderate the relationship between psychological contract breach and trust (T3), such that those with low prior trust should experience a greater decline in trust following a psychological contract breach than those with high prior trust. To test this hypothesis, moderated regression was performed. The interaction terms used in this approach are likely to correlate with the variables from which they were created. To reduce this multicollinearity, Aiken and West (1991) and Cronbach (1987) recommended that the independent variables first be centered around zero, subtracting each variable from its respective mean. This transformation does not affect the correlations among the variables, and it allows for a better estimate of the interaction term. In the first step, the centered dependent variable was regressed on the independent variable along with the control variables. In the second step, the interaction term was added to the equation.