4.1. Hypothesis tests
First, we tested the direct relationships between servant leadership and outcomes. Hypotheses 1–2 predicted servant leadership
would be positively associated with helping and creative behavior. As shown in Tables 2 and 3, both individual-level and
unit-level servant leadership were significantly associated with both helping and creative behavior in the predicted direction.
This supports Hypotheses 1 and 2.
Hypothesis 3 predicted that job satisfaction would mediate the relationship between servant leadership and patient satisfaction.
Results to support this hypothesis can be found in the moderated mediation results from Table 4 steps one (effect of servant
leadership on the mediator job satisfaction) and three (effect of servant leadership and job satisfaction on patient satisfaction).
Our results indicated that unit-level servant leadership was positively associated with job satisfaction, and job satisfaction was
positively associated with all three indicators of patient satisfaction. Therefore, Hypothesis 3 was supported.
Then we tested organizational structure as a moderator. Hypotheses 4–5 predicted that organizational structure would moderate
the relationships between servant leadership and nurse helping and creative behavior. For these individual-level outcomes,
we followed steps for testing cross-level interactions outlined by Aguinis, Gottfredson, and Culpepper (2013). As shown in
Tables 2 and 3, we found the interaction term of unit-level servant leadership and structure was significant only for creative
behavior1 (see Table 3). We followed recommendations by Aiken and West (1991) to graph the form of the significant interaction.
Results in Fig. 2 suggest unit-level servant leadership has a stronger positive relationship with creative behavior when structure is
high as opposed to when structure is low. A simple slopes test supported this interpretation, as the slope was non-significant
for low organizational structure, but the slope was significantly positive for moderate structure (.13, p b .01), high structure
(.23, p b .01) and very high structure (.32, p b .01). However, overall levels of creative behavior were higher for low structure regardless
of servant leadership. Thus, Hypothesis 5 was supported for unit-level servant leadership, but Hypothesis 4 was not
supported.
Finally, we tested Hypothesis 6 using moderated mediation. We analyzed structure as a moderator of the path from servant
leadership to job satisfaction analogous to what Edwards and Lambert (2007) call “first stage moderation” (p. 8). Because this
model tests the effect of servant leadership on job satisfaction as contingent upon structure, naturally then the indirect effect
of servant leadership on patient satisfaction is conditional on structure; hence the mediation effect is moderated by structure
(Hayes, 2012). We found support for this hypothesis as the conditional indirect effect of servant leadership on patient satisfaction
4.1. Hypothesis testsFirst, we tested the direct relationships between servant leadership and outcomes. Hypotheses 1–2 predicted servant leadershipwould be positively associated with helping and creative behavior. As shown in Tables 2 and 3, both individual-level andunit-level servant leadership were significantly associated with both helping and creative behavior in the predicted direction.This supports Hypotheses 1 and 2.Hypothesis 3 predicted that job satisfaction would mediate the relationship between servant leadership and patient satisfaction.Results to support this hypothesis can be found in the moderated mediation results from Table 4 steps one (effect of servantleadership on the mediator job satisfaction) and three (effect of servant leadership and job satisfaction on patient satisfaction).Our results indicated that unit-level servant leadership was positively associated with job satisfaction, and job satisfaction waspositively associated with all three indicators of patient satisfaction. Therefore, Hypothesis 3 was supported.Then we tested organizational structure as a moderator. Hypotheses 4–5 predicted that organizational structure would moderatethe relationships between servant leadership and nurse helping and creative behavior. For these individual-level outcomes,we followed steps for testing cross-level interactions outlined by Aguinis, Gottfredson, and Culpepper (2013). As shown inTables 2 and 3, we found the interaction term of unit-level servant leadership and structure was significant only for creativebehavior1 (see Table 3). We followed recommendations by Aiken and West (1991) to graph the form of the significant interaction.Results in Fig. 2 suggest unit-level servant leadership has a stronger positive relationship with creative behavior when structure ishigh as opposed to when structure is low. A simple slopes test supported this interpretation, as the slope was non-significantfor low organizational structure, but the slope was significantly positive for moderate structure (.13, p b .01), high structure(.23, p b .01) and very high structure (.32, p b .01). However, overall levels of creative behavior were higher for low structure regardlessof servant leadership. Thus, Hypothesis 5 was supported for unit-level servant leadership, but Hypothesis 4 was notsupported.Finally, we tested Hypothesis 6 using moderated mediation. We analyzed structure as a moderator of the path from servantleadership to job satisfaction analogous to what Edwards and Lambert (2007) call “first stage moderation” (p. 8). Because thismodel tests the effect of servant leadership on job satisfaction as contingent upon structure, naturally then the indirect effectof servant leadership on patient satisfaction is conditional on structure; hence the mediation effect is moderated by structure(Hayes, 2012). We found support for this hypothesis as the conditional indirect effect of servant leadership on patient satisfaction
การแปล กรุณารอสักครู่..
