2.2.6. Analyses
We used hierarchical regression analyses to test our hypotheses
about the two-way interaction effects of the two dimensions of
work stress, two forms of supervisory task feedback, and employee
creativity. The work stress and supervisory task feedback scales
were centered at their means before computing the interaction
analyses (Aiken and West, 1991). To enable any significant interaction
effects to be more closely examined, we plotted the simple
slope of the work stress and employee creativity regression at one
standard deviation above and below the mean of each item of positive
and negative task feedback from the supervisor, and tested
whether each slope was significant, according to Aiken and West’s
(1991) recommendation. In addition, we performed confirmatory
factor analyses (CFA) to investigate the discriminant validity of the
factor structures of work stress and task feedback from supervisors.
Overall model fit was assessed by the comparative fit index (CFI),