characteristics of this sample were, thus, very similar to the characteristics of the larger
sample of focal working individuals.
Moreover, I compared the means of the selected sample with the excluded sample
on the four culture dimensions to check If there were systematic differences between the
initial set of respondents, and the selected respondents. The differences between three of
the four culture dimensions were not significant. Only the means for market culture
differed significantly for both samples such that individuals used in the analysis reported
slightly higher values on market culture (F= .63, p< .05). This difference may suggest
that on average respondents in organizations higher on market culture perceive the
organizational level as more salient, thus, reporting culture on the organizational level.
The means of the other measures did not differ in the two samples.
Section two: relationships between roles and behaviors. For the second set of
analyses, the focus was on roles and behaviors. There were two sources used to measure
focal employee behaviors: coworkers and supervisors. Here a different procedure was
employed to select the most theoretically viable sample. For instance, for the coworker
sample, the focus was on retaining those coworkers who are familiar with the focal
individual behaviors such that their report would be most informed. The sample resulted
in 107 focal observations and 325 coworkers (average N=3 respondents per focal). In
this sample, the focal characteristics were very similar to the ones described in the
previous samples. The sample for supervisors was selected using a similar method by
focusing only on those supervisors who have reported knowing the person’s work well or
very well. This resulted in a sample of 143 supervisors responding for 89 focal
individuals (average N=1.66 per focal).