Abstract
A multitrait-multimethod matrix of leadership criteria, originally analyzed by the principal components procedure and rotated to varimax criterion, was reanalyzed employing Holzinger's bifactor analysis. In the original analysis only method factors were obtained, while the bifactor analysis isolated a trait factor in addition to method factors. The results are more meaningful and also give a better fit to the data, demonstrating the importance of a flexible approach in the choice of analytic procedures.