Strengthening Mediation-Based Management Research: Translational Value, Statistical Power, and Covariate Decisions

Abstract

The goal of my dissertation is to address the gap between research and practice in management scholarships. Despite decades of methodological innovation and theoretical advancement, many organizational research findings fail to translate into actionable guidance for practitioners. Conversely, questions faced by managers are not well represented in academic research. This disconnect enlarges the research-practice gap.To bridge this gap, my dissertation pursues two goals: 1. Advancing methodological rigor by investigating the statistical foundations of commonly used models in organizational research. 2. Improving practical applicability by clarifying the conditions under which these methods yield reliable, interpretable, and generalizable insights for both scholars and practitioners. At the core of this project is a simple but important set of research questions: • How much statistical power do researchers typically have when testing mediation models, which are among the most widely used approaches in organizational research? • How do design choices, such as sample size, covariates, and measurement assumptions, shape the validity and interpretability of mediation findings? • What implications do these methodological considerations have for bridging the gap between academic research and organizational practice? By addressing these questions, my dissertation aims to provide both researchers and practitioners with clearer guidance on when mediation models are sufficiently powered, when conclusions can be trusted, and how study design decisions affect both scientific rigor and practical relevance. The dissertation is organized into three papers, each addressing a different aspect of this problem: • Chapter 1 develops the conceptual foundation for understanding translation in management science. • Chapter 2 examines the statistical power of mediation models through large-scale simulations. • Chapter 3 investigates how covariates affect the power and interpretability of mediation findings. Together, these studies provide a systematic evaluation of mediation research in management, clarify the methodological conditions under which findings are robust, and offer practical recommendations for conducting and interpreting mediation analyses.