Leadership theories often include a contingency effect where the relationship between two or more variables is normally theorized to be monotonic, i.e., that it has a generally accepted direction—positive or negative—across the full range of the contingency variable. Most examinations of contingencies estimate how the monotonic relationship changes at mean, or near mean, levels of the moderator variable. We push the logic of moderation further to explore whether, for extra-ordinary values of the moderator, the effect may actually become non-monotonic, e.g., invert by moving from positive to negative in slope (or vice versa). Discovering such inversion effects in models of leadership would provide a deeper understanding of the operation and boundaries of theories, thereby calling for refinements of underlying theoretical assumptions. Using an innovative inductive approach, we search the leadership literature and find studies where extra-ordinary moderator values signal a potential inversion effect. We narrow onto two example leader-member exchange (LMX) studies to inductively theorize the mechanisms creating the inversion. We then generalize the logic of this mechanism to propose new theory for why such inversions might be occurring in a wider range of phenomenon beyond LMX, and discuss the associated implications for leadership and organizational theory.
CAVARRETTA, F., HANNAH, S.T., PICCOLO, R.F. et UHL-BIEN, M. (2015). Leadership beyond the Tipping Point: Toward the Discovery of Inversions and Complementary Hypotheses. ESSEC Business School.