Date of Conferral


Date of Award



The Multifactor Leadership Questionnaire (MLQ), the most used instrument for gauging Bass and Avolio's full-range leadership model, underwent numerous revisions to answer criticisms about its construct validity. Since researchers found conflicting results regarding the number of factors that best constitute the model, this study examined whether the factor structure of the MLQ (5X) was consistent across a diverse array of samples. The total size of the samples was 6,525, integrated from 18 independent studies. Using confirmatory structural equation modeling techniques, results indicated that the factor structure was best represented by nine single-order factors, as asserted by Bass and Avolio. These results were prevalent when all samples were integrated for the factor structure invariance test, or when individual samples were grouped into homogenous units for strict factorial or factor structure invariance. Thus, the validity of the model was a function of sample conditions explaining why independent researchers who used nonhomogenous samples failed to confirm the nine-factor model. These conditions---interpretable as moderators that bounded the theoretical model---included various environmental and organizational settings, leader gender, and the hierarchical level of the leader. The criterion validity of the MLQ was also tested, but results should be viewed cautiously as the leadership and outcome measures were each collected from the same source. Nonetheless, the regression paths of the MLQ factors were interpretable, and followed theoretical propositions. Transformational and contingent reward leadership was positively related to performance, and passive-avoidant leadership negatively related. Management-by-exception active varied according to sample conditions. These findings imply that a standard set of leadership behaviors is not universally related to effectiveness, and that leaders operationalize their behaviors differently depending on contextual factors. As a consequence, the interfactor relationships of the MLQ, and the relationships of the factors to criterion measures may be a function of the conditions under which the model is examined. Therefore, testing the MLQ's validity---and indeed that of other leadership instruments---should be performed under homogenous sample conditions. Finally, this dissertation suggests that the MLQ should be retained for future research and training. Possible improvements to the theory and measurement model are also discussed.