Natural Experiments in Management Studies

Identifying causal relationships is a difficult endeavour in management research. Conducting randomized experiments is difficult, whereas observational studies are plagued by endogeneity (Antonakis et al. 2010; Semadeni et al. 2014). However, management researchers increasingly realize that it is important to not just identify statistical relationships, but rather causal relationships. Natural Experiments represent an approach to overcome these difficulties. Natural experiments are situations in which an external factor (e.g., nature or governments) almost randomly assign a treatment to some subjects (treatment group) but not to others (control group). Yet, despite their potential to identify causal relationships, natural experiments are seldom used in management research.

On this website, you will find information about natural experiments and how they can be applied in management research. Particularly, you will find information regarding the three most frequently used designs: Difference-in-difference design, instrumental variable design, and regression discontinuity design. Additionally, you find some recommended readings if you would like to learn more about natural experiments and some information about helpful software that facilitate the analysis of natural experiments.