Replicability of Experimental Data and Credibility of Economic Theory

Abstract

Finding a data set that supports a theory is not enough for the theory to gain credibility. Credibility of a theory is established only after the initial supporting finding can be replicated by a number of follow-up studies. Economics so far has paid little attention to replicability, presumably because the profession has been overly pessimistic about obtaining a high degree of replicability of non-trivial theoretical predictions. By using a large data set I have collected that involves more than 4,000 subjects, I show that there is hope for economic theory to gain credibility by means of replications of laboratory data.