Measures of Authentic, Ethical, and Servant Leadership Spuriously Predict Objective Outcomes

Abstract

In two experiments, we provide evidence that questionnaire measures of authentic, ethical, and servant leadership are not valid representations of leader behaviors. We demonstrate that the measures are affected by irrelevant factors and can spuriously predict outcomes. In Study 1, a manipulation of performance cues and charismatic tactics impacted ratings of these leadership measures. In Study 2, ratings of leadership style measures predicted an objective outcome, even when leader behaviors did not vary and when we controlled for performance cues and rater-specific variables. This latter finding suggests that these leadership styles measure more than leader behaviors, although researchers use these styles as if they were behavioral constructs. We speculate that leadership styles as constructs-in-use differ from leadership styles as constructs-as-theorized. Whereas leadership styles as constructs-in-use appear to confound leader behaviors and evaluations of these behaviors, theories of leadership styles treat these styles as patterns of leader behaviors. Consequently, research suggesting that authentic, ethical, and servant-like leader behaviors engender positive results might be a mere artifact of poor measurement. Thus, future research should only use questionnaire measures of leadership styles if these questionnaires actually reflect the construct-as-theorized.