2002), the direct measure proved to have greater utility for my outcomes of interest.
Based on my results, I’d recommend that construct validation would be an important
aspect of using indirect measures of strength instead of assuming the meaning that they
carry. According to the correlation tables from two of the sub-samples I used, the
standard deviations were mostly not significantly related to the direct measures of
strength (correlations ranging from .01 to -.28).
Only in the case of market culture and hierarchy culture, the standard deviations
were negatively related to the respective measures of strength (-.28, p< .05 & -.24, p<
.05), suggesting that the level of disagreement on market culture had a negative
association with a direct measure of culture strength on the same dimension. However,
there is also a positive association between the standard deviation on entrepreneurial
culture and entrepreneurial culture strength (.29, p< .05), in contrast to what would be
theoretically expected. There may be several theoretically viable explanations of the
observed relationships. Specifically, it is possible that the high standard deviation on
entrepreneurial culture reflects the presence of very high and very low scores on
entrepreneurial culture. Hypothetically, if the focal employees who provided the direct
measure of strength were consistently in agreement with the coworkers who provided the
higher ratings on entrepreneurial culture, this tendency would be reflected in a positive
correlation between the standard deviation and the direct measure of strength. In addition,
the low number of coworkers who provided ratings of culture resulted in lower reliability
of their ratings.
Given the overall inconsistency in the findings with respect to strength as a
standard deviation and strength as agreement, it seems warranted for future research in