BUS603: Managing People

Preliminary analysis

First, a confirmatory factor analysis was conducted on four variables established in this study, and a further analysis was conducted excluding three items with low factor loadings of authentic leadership. The results and questionnaire items are shown in Table 1. The incremental fit index (IFI) and comparative fit index (CFI), which argued must be reported in order to determine the goodness of fit of the model, both exceed 0.90. In addition, the measurement variables are loaded significantly on variables as originally intended, thereby securing the construct validity of variables. Cronbach's alpha was measured by conducting a reliability analysis based on the above. The results show that authentic leadership is 0.92, relational cohesion is 0.96, hedonic well-being is 0.92, and eudaimonic well-being is 0.95, satisfying the condition that reliability is secured when it is at least 0.70 in general.

Prior to the actual testing of the hypotheses, a correlation analysis was conducted to determine the relevance among variables. The correlation among control variables shows that male and married employees have high positive correlations with hedonic well-being/eudaimonic well-being. Moreover, college graduates or higher show a positive correlation with hedonic well-being. Individualism and collectivism also showed a positive correlation with hedonic well-being/eudaimonic well-being. Above all, transformational leadership and ethical leadership show strong positive correlations with the key variables, increasing their validity as control variables. The correlation among the key variables shows that authentic leadership has strong positive correlations with hedonic well-being (0.246, p < 0.01) and eudaimonic well-being (0.313, p < 0.01). Relational cohesion (a moderating variable) also shows strong positive correlations with hedonic well-being (0.381, p < 0.01) and eudaimonic well-being (0.387, p < 0.01). These results suggest the validity of the hypotheses. Table 2 presents the results of the correlation analysis.

Table 2.

Means, standard deviations, and correlations between the key variables (n = 950)
* p < 0.05, ** p < 0.01 (two-tailed).
Industry 1 = manufacturer; others = 0 dummy variables.
Industry 2 = public enterprise; others = 0 dummy variables.
General employee, male, married, and college graduate = 1; others = 0 dummy variables.

Test of hypotheses

A hierarchical regression analysis was used to test the hypotheses. The results are presented in Table 3. Model 1 shows the effects of control variables. College graduates or higher and employees with a strong collective disposition tend to perceive higher hedonic well-being (respectively, β = 0.26, p < 0.05; β = 0.30, p < 0.001). Moreover, employees in manufacturing show lower hedonic well-being and eudaimonic well-being compared with those in finance and public enterprises, indicating that there are differences among industries (respectively, β = −0.31, p < 0.01; β = −0.29, p < 0.01). In addition, employees with high individualism and collectivism perceive higher eudaimonic well-being (respectively, β = 0.12, p < 0.001; β = 0.28, p < 0.001). Lastly, the effects of ethical leadership on well-being are not significant; however, the direct effects of transformational leadership on hedonic and eudaimonic well-being are strong, thereby showing validity as a control variable (respectively, β = 0.28, p < 0.001; β = 0.30, p < 0.001).

Table 3.

Result of hierarchical regression analysis (n = 950).
† p < 0.10, * p < 0.05, ** p < 0.01, and *** p < 0.001 (two-tailed).
Industry 1 = manufacturer; others = 0 dummy variables.
Industry 2 = public enterprise; others = 0 dummy variables.
General employee, male, married, and college graduate = 1; others = 0 dummy variables.

As presented in Model 2, which shows the direct effects of authentic leadership, these effects on hedonic well-being are not significant, thereby rejecting Hypothesis 1 (β = −0.03, ns). However, as expected, authentic leadership has positive effects on eudaimonic well-being, thereby confirming Hypothesis 2 (β = 0.10, p < 0.10). Model 3 shows the effects of relational cohesion. These effects on hedonic and eudaimonic well-being are both significant (respectively, β = 0.26, p < 0.001; β = 0.22, p < 0.001).

Finally, Model 4 verifies the moderating effect of relational cohesion. Here, the interaction effect on hedonic well-being is positive (β = 0.04, p < 0.10). In other words, although the direct effects of authentic leadership on hedonic well-being are not significant, if relational cohesion is highly perceived, as shown in the graphs of Figure 1, the effects of authentic leadership are positive. Likewise, the interaction effect on eudaimonic well-being is also positive, thereby proving that the moderating effect of relational cohesion is significant (β = 0.10, p < 0.001).