This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: 1 The Role of Theory in Field Experiments 1 David Card, Stefano DellaVigna, Ulrike Malmendier UC Berkeley and NBER February 8, 2011 Preliminary Draft for JEP special issue on Field Experiments I. Introduction When it comes to role of theory, empirical microeconomists are torn. On the one hand, we devote a large fraction of our graduate courses to models of consumer and firm deci- sion-making, and to the interactions that determine market equilibrium. On the other hand, it is not always obvious how these theories are relevant to empirical research. Out- side the academy, policy-makers and business leaders often demand basic facts and simplified policy guidance with little or no concern for theoretical nuances. How then do empirical economists negotiate between the demands for theory and facts? In this paper, we focus on the role of theory in the rapidly-growing area of field experiments. We take an empirical approach and quantify the role of theoretical modeling in all published field experiments in the top five economics journals from 1975 to 2010. Specifically, we propose a new classification of experimental studies that captures the extent to which the experimental design is linked to economic theory. We distinguish be- tween four types of studies: studies that do not contain an explicit model ( Descriptive ); studies that test a single model-based hypothesis ( Single Model ); studies that test compet- ing model-based hypotheses ( Competing Models ); and studies that estimate structural pa- rameters in a completely specified model ( Parameter Estimation ). Applying the same classification to laboratory experiments, we show that theory has played a more important role in the laboratory than in field experiments. We discuss in detail three sets of field experiments that illustrate the potential promise and the pitfalls of a tight link between experimental design and theoretical underpinnings. II. Quantifying the Role of Theory in Field Experiments 1 We thank for helpful comments the editors (David Autor and John List), as well as Iwan Barankay, Glenn Harrison, Matthew Rabin, David Reiley, and participants in Berkeley, at the Wharton Conference on Field Experiment, and at the 2011 ASSA conference in Denver. We thank Ivan Balbuzanov and Xiaoyu Xia for excellent research assistance. 2 a. Defining Field Experiments The use of experimental techniques -- random assignment of a manipulation of interest -- came relatively late to economics. 2 In the last 15 years, however, randomized experi- ments conducted in field settings have become more common, and in 2010 field experi- ments represent about 3 percent of the articles published in the top journals. As in other areas of applied economics, the role of theory ranges from almost none to fully model- based investigations. Nevertheless, there is a widespread perception that many experi- mental studies particularly field-based random-assignment studies are black box evaluations that provide only limited evidence on theoretically-relevant mechanisms (see,...
View Full Document
- Fall '08
- The American