Quasi and Natural Experiments in Public Management Research

• In Event: Causal Inference in Public Management

Abstract

For much of its history, public management research relied primarily on observational studies, such as surveys or analyses of administrative data, that provided only weak evidence about theoretically important cause-effect relationships. In response, the last decade has seen a surge in the use of randomized experiments in public management research, including a wide variety of survey experiments as well as some lab and field experiments. Indeed, many top public management journals now prefer randomized experiments over observational studies. 

But an important middle ground has been overlooked in public management research, namely, the use of quasi and natural experiments. This is a somewhat curious oversight, given that much public management theory and research focuses on real-world organizations and practices that often are not amenable to full random assignment. In many such applied contexts, looking for natural experiments or creating quasi-experiments provides a widely useful alternative to fully randomized studies. Based on a systematic review of causal study designs in the top three journals in the field (PAR, JPART, PMR) over the past 20 years, we find that, among randomized experiments, survey experiments predominate (n=223), with fewer field experiments (n=47) and lab experiments (n=18). In contrast, quasi and natural experiments are relatively uncommon (n=33). Despite their limited number, we highlight some important examples of quasi and natural experiments that illustrate their potential for public management research.

We next examine the public management literature using a new framework that describes three key dimensions of randomized, quasi, and natural experiments (Remler & Van Ryzin, 2025) and provides a road map for both consumers and producers of causal evidence. For example, the framework highlights that both researcher-controlled and natural experiments can be used for a wide variety of randomized as well as quasi-random assignment strategies. The framework also highlights how both randomized and quasi-random designs can be used for indirect studies, such as encouragement experiments. Applying the framework to the articles from the systematic review, we identify overlooked study types, such as indirect natural experiments and researcher-controlled quasi-experiments. We use this framework to illustrate how a wide variety of quasi and natural experiments could broaden the scope of well-estimated causal effects of interest to public management theory and practice.

Remler, D. K., & Van Ryzin, G. G. (2025). Control, exogeneity, and directness: Understanding and designing quasi and natural experiments. American Journal of Evaluation, 46(1), 62-89.

Authors

• Gregg G Van Ryzin, Rutgers University - Newark
  Presenting Author

• Dahlia Remler, Baruch College (City University of New York)
  Non-Presenting Co-Author