Internal and External Validity in Economics Research: Trade-offs between Experiments, Field Experiments, Natural Experiments a

In the realm of empirical research, investigators are first and foremost concerned with the validity of their results, but validity is a multi-dimensional ideal. In this article we discuss two key dimensions of validity – internal and external validity – and underscore the natural tension that arises in choosing a research approach to maximize both types of validity. We propose that the most common approaches to empirical research – the use of naturally-occurring field/market data and the use of laboratory experiments – fall on the ends of a spectrum of research approaches, and that the interior of this spectrum includes intermediary approaches such as field experiments and natural experiments. Furthermore, we argue that choosing between lab experiments and field data usually requires a tradeoff between the pursuit of internal and external validity. Movements toward the interior of the spectrum can often ease the tension between internal and external validity but are also accompanied by other important limitations, such as less control over subject matter or topic areas and a reduced ability for others to replicate research. Finally, we highlight recent attempts to modify and mix research approaches in a way that eases the natural conflict between internal and external validity and discuss if employing multiple methods leads to economies of scope in research costs. *Roe is professor in the Department of Agricultural, Environmental and Development Economics, Ohio State University. Just is an associate professor in the Department of Applied Economics and Management, Cornell University. Roe acknowledges support of the Ohio Agricultural Research and Development Center. This article was presented in an invited paper session at the 2009 ASSA annual meeting in San Francisco, CA. The articles in these sessions are not subjected to the journal‟s standard refereeing process. The authors thank Stephan Marette, Ernst Fehr and Rachel Croson for helpful suggestions.