Repeated measures regression and mediational tests: Enhancing the power of leadership research

Abstract

Repeated measures regression is described and advocated for team contexts in which mediational tests are appropriate. We propose that this paradigm serves as an excellent design and data-analytic strategy for research on teams and team leadership, because studies in this area are often characterized by a small number of teams and leaders, relative to the number of variables that might be theoretically linked to their outcomes. The technique is especially valuable for longitudinal studies, where practical limitations preclude a large number of teams. Although the longitudinal nature of such research often leads to small sample sizes and hence reduced power, repeated measures regression makes it possible to regain some of this power through the use of multiple observations and between teams/within team variance partitioning. The use of this technique is illustrated with both hypothetical data and actual data generated to test a specific theory of teams and team leadership.