Observing Leadership as and in Networks: A Multilevel Investigation of Shared Leadership, Discrete Emotions, and Knowledge Behaviors

PURPOSE: With an increasing demand for shared leadership to address complex, dynamic, and diverse knowledge situations, more attention should be paid to the knowledge behaviors of emergent leaders in teams. However, there is thus far a lack of research into the multilevel nature of shared leadership. Using a dual network lens, this study considers two aspects of shared leadership: “shared leadership as networks” (SLAN) and “shared leadership in networks” (SLIN). Based on emotion appraisal theory, this study investigated the impact of SLIN on leaders’ productive and counterproductive knowledge behaviors through discrete emotions (pride and fear of losing power) and the moderating role of SLAN in guiding their emotions and subsequent behaviors. METHODS: A social network approach was employed to obtain data among 431 employees from 72 teams in China. We identified 350 emergent leaders via leadership sociograms. Statistical analysis methods, including multilevel regression analysis, Monte Carlo simulation, and simple slope analysis, were utilized to test our hypotheses. RESULTS: This study indicates that (1) SLIN is positively associated with pride (B = 0.33, p < 0.001), which in turn increases knowledge sharing (B = 0.37, p < 0.001); (2) SLIN is positively related to fear of losing power (B = 0.21, p < 0.05), which in turn is negatively associated with knowledge sharing (B = −0.23, p < 0.001) and positively associated with knowledge hiding (B = 0.19, p < 0.001); (3) SLAN moderated these aforementioned indirect effects. CONCLUSION AND ORIGINALITY: First, by defining and differentiating between SLAN and SLIN, this study expands a new perspective on SLIN and facilitates the further development of multilevel shared leadership. Second, this study responds to recent calls for more research on leadership networks and individual-level outcomes of shared leadership. Moreover, it offers a novel theoretical framework to illustrate the impact of SLIN on their knowledge behaviors. Finally, the focus on the moderating role of SLAN identifies a significant contextual factor that prevents the dysfunctional effects of SLIN. These original discoveries enrich the literature on team leadership, emotions, and knowledge management. Further, our novel findings may be of interest to practitioners as they help organizations and teams manage the complex impacts of SLIN on knowledge behaviors in the contexts of an ongoing “affective revolution” and team-based organizational structures.