Subcoalition Cluster Analysis: A New Method for Modeling Conflict in Organizations.

This article introduces Subcoalition Cluster Analysis (SCA), a novel computational framework designed to quantitatively model business firms as political systems composed of latent subgroups called "subcoalitions," which represent clusters of actors with internally aligned but externally conflicting preferences. SCA addresses the methodological gap in studying intrafirm conflict by partitioning heterogeneous actors into stable subcoalitions based on their pairwise preference data, using an optimization approach related to—but computationally more tractable than—classical preference aggregation methods like Kemeny's method. The paper demonstrates SCA’s applicability through two empirical cases: Wikipedia’s editor community, revealing a split between experienced editors concerned with cultural issues and newer editors focused on technical problems; and the Baseball Writers' Association of America, showing the emergence and persistence of two subcoalitions divided over the consideration of suspected performance-enhancing drug users for Hall of Fame induction. The authors discuss SCA’s theoretical foundations, algorithmic structure, methods for determining the optimal number of subcoalitions, and potential for future research in organizational conflict, emphasizing its utility for management researchers analyzing complex stakeholder governance and intrafirm political dynamics.