Leadership is a multilevel phenomenon. Accordingly, levels-of-analysis issues should be considered and addressed in theoretical and empirical studies of leadership. This article discusses levels-of-analysis issues in the conceptualization and measurement of leadership constructs, levels-of-analysis issues in developing leadership theories, change-related issues in leadership research, and up-to-date statistical methods for testing multilevel leadership theories. In this articles discussions, levels-of-analysis issues are illustrated by reviewing leadership literature and suggesting possible ways to further contribute to the understanding of leadership process in organizations. The authors hope this article can facilitate more rigorous leadership studies, help integrate different leadership theories, and inspire new approaches to leadership research.
