Multiverse Analysis to Advance Theory and Practice: Illustrating and Investigating Variability in CEO Gender Pay Inequality

Despite using valid research designs informed by substantive and methodological theory, results across studies fail to converge at times. We introduce multiverse analysis as a tool to help catalyze theory development in situations with divergent theoretical explanations and associated evidence by (a) offering strong and precise inference testing, (b) identifying sources of variability across (dozens, hundreds, thousands, or even millions of) systematically and comprehensively derived valid alternative design decisions, and (c) allowing for the multifaceted nature of theoretical mechanisms. We use the theoretical and empirical divergence surrounding studies on how chief executive officer (CEO) gender relates to CEO pay to illustrate how multiverse analysis is a systematic and comprehensive method that is widely applicable, accessible, and flexible, making it a valuable catalyst for theory building.