In addition, our findings also point out that seniority is positively related to resignation. This particular result confirms the findings of previous research showing the positiverelationship between continuance commitment and job tenure Because this form of commitment refers back to the costs that may be incurred by employees as a result of a decision to leave the organization, this concept is very close to that of resigned satisfaction. This finding suggests that the longer an individual stays in their organization, the more likely they will be resigned. These results are much like those of previous studies that point out that the relationship between job tenure and a decreased level of job satisfaction may be explained by the ‘accepted’ presence of ‘red tape’ . Considering step 2 of the same hierarchical regression analysis (table 3), it is abundantly clear that red tape is the strongest predictor of resignation. Thus, the explained variance of the dependent variable increased noticeably (step 1: R2 = .094; step 2 : R2 = .175;_R2 = .081). This result clearly supports hypothesis 1. In previous research, red tape has been reported as a strong predictor of work dissatisfaction in the public sector. In this respect, our own research confirms that research.
The results of the second hierarchical regression analysis (table 4) show that the different dimensions of PSM have divergent effects on resignation. The addition of the PSM dimensions to the control variables (step 2 of the second hierarchical regression analysis, presented in table 4) has only a little effect on the explained variance of resignation (step 1: R2 = .094; step 2 : R2 = .138;_R2 = .044). However, this model clearly demonstrates that the four dimensions of PSM do not have an equal effect on resignation. The ‘attraction to policy-making’ and ‘commitment to public interest/civic duty’ PSM dimensions have a negative and significant effect on resignation, whereas the ‘compassion’ and ‘self-sacrifice’ PSM dimensions have a positive and significant effect on resignation. As stated by Taylor,‘the negative association between affective motives and job motivation may suggest that the respondents perceived their work environment as being less than supportive in helping them achieve their desire to serve the public’. These results partially contradict other studies pointing out the positive effect of ‘self-sacrifice’ on job satisfaction. However, our results confirm that the different public service orientations encompassing the PSM construct do have divergent effects on employees’ outcomes, such as job satisfaction, and, for the specific purpose of this article, resignation. In other words, the different orientations towards public service have different impacts on the equilibrium between individual expectations or aspirations and work situation. Our results suggest that ‘attraction to policy-making’ and, to a larger extent, ‘commitment to public interest/civic duty’ represent the PSM dimensions that are more compatible with the realities of bureaucratic work.
However, we should underline the positive association between the ‘compassion’ and the ‘self-sacrifice’ PSM dimensions and resignation. This result seems to suggest that respondents driven by these kinds of public service orientations may consider that their work conditions do not allow them to fulfil their needs and expectations. Thus, it is very likely that these individuals perceive a stronger incompatibility between their individual
aspirations and their work situation. Consequently, with respect to the above results, hypotheses 2 and 3 are supported in the context of our study. The last hierarchical regression analysis (table 5) containing the control variable, red tape, the dimensions of PSM, and the new variables created to measure the moderating effect of PSM dimensions on the relationship between red tape and resignation, accounted for 21.8 per cent of the variance of resignation. The four models presented in table 5 display two general findings. First, red tape remains the strongest predictor of resignation, even when taking into account the control (step 1) and the other independent variables (step3 and 4). Secondly, the dimensions of PSM have differentiated impacts on resignation while taking into account the other variables in the model. Three out of the four PSM dimensions are significantly related to resignation, but their effects do not converge.
‘Compassion’ is positively but weakly related to resignation. The same is true for ‘selfsacrifice’,but the relation is not statistically significant. Conversely, ‘commitment topublic interest/civic duty’ and, to a lesser extent, ‘attraction to policy-making’ havea negative impact on resignation. Finally, an additional but weak part of variance is accounted for when introducing the PSM dimensions to the model of regression (step 1: R2 = .094; step 2: R2 = .175; step 3: R2 = .213;_R2 = .038). This last result confirms that PSM dime