X hits on this document





133 / 185

The other proposed interaction effects did not reach statistical significance

(although some of them were close to the .05 threshold). For instance, the interaction

term between clan and perceived fit had a negative influence on achievement role

perceptions and was close to reaching statistical significance (p < .10). The negative

coefficient suggests that when both perceived fit and clan culture are high or low (in the

presence of non-significant main effects for the interaction step), individuals tend to

perceive high achievement as less expected of their role in the organization, thereby

supporting the logic outlined in hypothesis 7b.

Relationships between Roles and Behavior

The second part of the model examined roles as predictors of behavior and

possible boundary conditions (Biddle, 1979). I conducted analysis with both coworkers

and supervisors as a source. The logic for using coworkers was that they are more likely

to be able to directly observe the everyday activities, and behaviors of the focal person

than supervisors (Murphy & Cleveland, 1990; Pollack & Pollack, 1996). Surprisingly,

very few of the proposed relationships were realized. Specifically, innovative role had a

strong negative relationship with compliant behaviors (coworker perspective), and both

compliant and innovative roles had links in the predicted direction with compliance and

innovation but they did not reach statistical significance. Overall, this suggests a weak

pattern of results in relating roles to behaviors, which may be rooted in under-specified

theory or in empirical problems.

From a theoretical perspective, the lack of relationships would suggest the

presence of moderators that act as boundary conditions determining the extent to which

individuals enact expected roles. From an empirical point of view, there may be many


Document info
Document views758
Page views758
Page last viewedMon Jan 23 23:49:57 UTC 2017