Dyads in Organizational Research: Conceptual Issues and Multilevel Analyses

Although extensive work exists on multilevel theory building, testing and inference drawing in organizational research, work focusing on interpersonal relationships and the dyad as a level of analysis is scarcer compared to research focusing on individuals, groups/teams, and organizations. This scarcity is attributable to conceptual and methodological challenges in modeling dyads as a level of analysis. To address these challenges, we first present a discussion of the conceptualization of different dyadic configurations in work settings. Then, drawing on this conceptual treatment, we develop and illustrate the application of Random Coefficient Modeling (RCM) via nesting for independent dyads and cross-classification for dependent dyads in work settings. Across both independent and dependent dyads, we illustrate the application of Within and Between Analysis (WABA), another multilevel technique for dyadic analyses. RCM and WABA stem from differing statistical assumptions and are capable of testing different kinds of multilevel questions. We close with prescriptive implications for future multilevel conceptualization and data analysis using RCM and WABA, particularly involving interpersonal work relationships and dyads.