Field Experiments and the Rubin Causal Model: Review of Approaches and Current Research

Abstract

Experiments of various kinds are increasingly being used in the social sciences to derive causal inference. Among the varieties of this method, field experiments are especially noteworthy. Explosive growth in their numbers has been observed in recent years, primarily in economics and political science. Gradually, field experimentation is starting to spread to other disciplines. One of the most important reasons for this is the popu­larization of the so-called Donald Rubin model of causal inference, which allows researchers to link experimental methods with statistics and other mathematical methods. In the Russian-speaking academic field, one can observe a lack of texts describing how field experiments are related to this model in causal inference, while such a research design allows us to focus specifically on the search for the causality of various social phe­nomena. This article provides a critical-bibliographic review of both the conceptual model of causation and the existing research carried out in the design of field experiments in the Rubin model. The first part of the paper provides a brief overview of the main paradigms of causation and how, from one of them (the approach of potential outcomes and counterfactual inference), the Rubin model logically arises. The following describes the milestones in the history of field ex­periments before the Rubin model. This is followed by a description of the model and today's debate about the advantages, limitations, and design features of the field experiment. Finally, with a few examples, we analyze several well-known field experiments to illustrate the operation of the described method.