20176: The “Nebula” of Leadership: A Review of Leadership Literature Based on Large Language Models (session 25548)

In the past century, leadership research has accumulated a vast array of constructs and theories. This study proposes leadership literature as a “nebula,” capturing its richness while emphasizing its ambiguities due to construct redundancy. Using a bibliometric analysis of 2,226 articles published between 2000 and 2023 with the application of large language models (LLM), this study identifies 146 primary leadership constructs and 389 leadership behaviors. We use LLM to analyze the proportionate distributions of leadership behaviors within leadership constructs and calculate similarities among leadership constructs. Our findings suggest a prevalence of construct similarity and redundancy, concentrated in clusters of “good” and “bad” leadership constructs. “Good” leadership constructs, while not directly integrable, can be synthesized by isolating their shared core components and incorporating the unique deviations, which include five sub-facets: promoting values and morality, driving organizational change and adaptability, emphasizing relationships and well-being, fostering cooperation, and focusing on task advancement. In contrast, “bad” leadership constructs can be integrated into two categories: Dictator and Saboteur. Overall, by reviewing existing constructs and identifying areas of redundancy, we provide a roadmap for advancing leadership research with greater conceptual clarity and parsimony.

view paper (if available)