Document Type



Civil Rights and Discrimination | Labor and Employment Law


Researchers often rely on published opinions to draw conclusions about cases decided by the courts, determinants of court decisions, and broader social phenomena. We demonstrate that 80 to 90 percent of employment discrimination cases filed in federal court do not produce a published opinion. There are good theoretical reasons to believe that the process generating a published opinion is not random and thus that samples of published cases will not be representative of all cases. Through a direct comparison of published and unpublished cases, we show that the two actually do differ in significant and predictable ways. Examining several studies that use cases with published opinions for a variety of purposes, we show how our understanding of the operation of employment discrimination law changes-in some instances, dramatically-when we look at all cases, whether or not they have published opinions.