The forgotten follower: a contingency model of leadership and follower self‐leadership
研究通过纵向数据和多层线性模型，发现授权型与指令型领导对追随者自我领导的影响取决于追随者的自主需求，为领导力权变理论提供了实证支持。
Purpose Seeks to examine the interaction effect of leadership and follower characteristics on follower self‐leadership, using hierarchical linear modeling. Design/methodology/approach Longitudinal data were collected using a questionnaire at two points in time, with ten weeks between each collection. These data facilitate the causal inference between leadership and follower need for autonomy (wave 1) and follower self‐leadership behaviors (wave 2). Hierarchical linear modeling (HLM) was used to analyze the hierarchical structure data. Findings Both empowering and directive leadership (group level) interacted with follower's need for autonomy (individual level) to enhance subsequent follower self‐leadership (individual level). That is, empowering leadership had a stronger positive effect on followers who were high on the need for autonomy, and directive leadership had a stronger negative effect on followers who were high on the need for autonomy. In summary, the influence of leadership on follower self‐leadership was contingent on follower need for autonomy. Overall, the results supported the view that attributes of the follower can be an important element in contingency theories of leadership. Research limitation/implications This study does not include other possible individual characteristics, group level characteristics, and organizational level or environmental characteristics. A future research design might include organizational‐level characteristics. Practical implications Both the leadership context and the trait of the individual employee work hand in hand to produce true self‐leadership. Therefore, organizations need to develop empowering leaders who will, in turn, develop followers who are effective at self‐leadership. Originality/value This research contributes to the literature by testing a contingency model of leadership and follower self‐leadership. This study also demonstrated the usefulness of HLM to test interaction effects between group‐level variables and an individual‐level variable on individual‐level dependent variables.