Natural Experiments in Business Research Methods

Causal identification is an important consideration for organizational researchers as they attempt to develop a theoretical understanding of the causes and effects of organizational phenomena. Without valid causal identification, insights regarding organizational phenomena are challenging given their inherent complexity. In other words, organizational research will be limited in its scientific progression. Randomized controlled experiments are often suggested to provide the ideal study design necessary to address potential confounding effects and isolate true causal relationships. Nevertheless, only a few research questions lend themselves to this study design. In particular, the full randomization of subjects in the treatment and control group may not be possible due to the empirical constraints. Within the strategic management area, for example, scholars often use secondary data to examine research questions related to competitive advantage and firm performance. Natural experiments are increasingly recognized as a viable approach to identify causal relationships without true random assignment. Natural experiments leverage external sources of variation to isolate causal effects and avoid potentially confounding influences that often arise in observational data. Natural experiments require two key assumptions—the as-if random assignment assumption and the stable unit treatment value assumption. When these assumptions are met, natural experiments can be an important methodological approach for advancing causal understanding of organizational phenomena.