METAANALYTIC STUDIES OF LEADERSHIP - ADDED INSIGHTS OR ADDED PARADOXES

This research examines the extent to which meta-analysis of leadership research fulfills its promise to bring order into a body of literatur e characterized by inconsistent, equivocal findings. Manual and comput er searches yielded 14 meta-analyses which clustered around the follow ing leadership topics: (1) theories of leadership (n = 7); (2) leaders hip in small groups (n = 2); and gender differences in leadership (n = 5). In all three categories, the results of the meta-analyses were as inconsistent as those of primary research. Differences in the types o f meta-analystic techniques used, differences in judgment calls includ ing decisions defining the research domain, establishing criteria for inclusion of studies, coding characteristics and selection of potentia l moderators are, at least in part, responsible for the differences ob served. In addition, areas of leadership research which have produced sizable bodies of empirical studies such as the effects of power on le adership outcomes, leader effectiveness as well as the effects of lead ership competencies and leadership contexts on leader-follower relatio ns have not been subjected to statistical aggregation and summarizatio n of results across studies. Recommendations for improving the quality of meta-analytic research on leadership are offered.