What Is Known about Testing for Discrimination: Lessons Learned by Comparing across Different Markets

The paper provides a fairly comprehensive examination of recent empirical work on discrimination within economics. The three major analytical approaches considered are traditional regression analysis of outcomes, paired testing or audits, and finally analysis of performance where higher group performance suggests that a group has been treated disfavorably. The review covers research in the labor, credit, and consumption markets, as well as recent studies of discrimination within the legal system. The review suggests that the validity of interpreting observed racial differences as discrimination depends heavily on whether the analysis is based on a sample that is representative of a population of individuals or households or based on a sample of market transactions, as well as the analyst?s ability to control for heterogeneity within that sample. Heterogeneous firm behavior and differentiated products, such as those found in labor and housing markets, also can confound empirical analyses of discrimination by confusing the allocation of individuals across firms or products with disparate treatment or by ignoring disparate impacts that might arise based on that allocation. An earlier version of this paper was sponsored by the National Resource Council of the National Academy and presented to the Panel on Methods for Assessing Discrimination on July 2, 2002. Any opinions, findings, conclusions, or recommendations expressed in this article are those of the author and do not necessarily reflect the views of the panel, the National Academy, or any agency of the U.S. Government.