UBC Theses and Dissertations
A meta-analysis of Fiedler’s Contingency model of leadership effectiveness Crehan, Ellen Patricia
Fiedler's Contingency Model of Leadership Effectiveness is widely cited, yet highly controversial. The present study subjected Fiedler-based studies to a meta-analysis to determine whether a body of consistent findings would emerge. If such findings did emerge, the aim would be to develop a theoretical framework which would adequately explain them. If they did not, the study would try to explain their absence. The original pool of 402 documents contained 112 primary studies. Of these studies, 38 met the criteria of quality, comparability, and extent of adherence to Fiedler's Model needed for retention in the meta-analysis. These 38 studies contained 249 LPC-performance correlations. These correlations were first analyzed for frequency of support for Fiedler's Model using two indicators: (1) the primary study authors' stated conclusions and (2) directional congruency with Fiedler's predictions. Full or partial support was concluded by 15% and 63%, respectively, of the authors. Directional congruency yielded 54% support and 46% non-support. The second phase of the analysis, in which the within-octant correlations were partitioned by eight different variables, examined both the magnitude and direction using median correlations as the common metric. The extensive disparity between the obtained medians and those predicted by Fiedler led to a third examination using Exploratory Data Analysis techniques. This analysis showed that, regardless of the variable used, there were within-octant differences which prevented not only combining the partitioned data sets but also continuing the meta-analysis. An analysis of the methodological variability in the studies yielded findings important to the continued use of Fiedler's Model. These findings led to several recommendations intended to standardize testing procedures. These recommendations included suggestions regarding greater standardization of score division methods for LPC and GAS, assessment of the three situational variables, and criteria by which to assess support for the Model. Until such standardization is achieved, the validity of Fiedler's Model can be neither confirmed nor disconfirmed.
Item Citations and Data