EPA→DSB path from the model and a return to SEM 5b:2 as the optimal model
(see Figure 7 - 22 ).
7.9.2 Model 5 / Model 5b summary
Model 5b has established that the strongest effects on job performance come
from motivational leadership (0.315), followed by employees’ positive attitudes
(0.279) and then from empowerment (0.174).
The development of the final model (5b) has seen the removal of the WV factor
from the model. This decision was made on the grounds that the WV→EPA path
was found to be non-significant and, in seeking a parsimonious solution (see e.g.
Byrne 2010: 185), this path was removed from the model. As noted above,
removing this path was tantamount to removing the WV factor entirely, thus the
structure for Model 5b (with no WV construct) was arrived at.
Models 5 and 5b represent alternative explanations of the same set of inter-
linkages and a comparison of these competing models was undertaken using the
range of model fit statistics described in Table 7 - 31. Following this comparison,
Model 5b was selected as the optimal explanation of the relationships between the
constructs.
7.10 Model
Model 6 sees the introduction of the Social Support (SS) construct to the model.
Social support has been included in three previous hospitality leadership studies
(Borchgrevink and Boster 1994; Ross and Boles 1994; Susskind et al. 2000a)
with mixed evidence for its positive influence in hospitality work environments
(see Section 5.2 for details). Within the organising framework for this research,
Social Support is hypothesised to be a predictor of Job Performance and Service
Quality.
Indicator variable SS1 was removed owing to a poor factor loading of 0.517 and
the remainder of the CFA model was developed along similar lines to previous
ones (i.e. the same factor structure as was reported for CFA 5:8b was resolved).
At this stage the model fit was adequate (χ^2 = 167.306, d.f. = 137, sig = 0.040;