Obsessive compulsive disorder (OCD) is increasingly studied in nonpatients, primarily through the selection of individuals who score high on a self-report measure of OCD. The usefulness of this methodology for understanding OCD presupposes that some of the individuals in the high-scoring group meet diagnostic criteria for OCD, that the obsessive-compulsive behaviors in the high-scoring individuals are stable across time to a certain degree, and that the features associated with OCD in patients also are found in the high-scoring nonpatients. Two studies are reported which provide support for these three assumptions. Together the studies suggest that OCD can be productively examined by the selection of individuals who score high on a self-report measure of OCD. Cautions in the use of this methodology for the study of OCD are also noted.