Since its introduction into the statistical genetics literature, the transmission disequilibrium test (TDT) has seen widespread use in analyses of linkage and association due not only to its simplicity but also to its desirable properties relative to other within-family analytic methods. In this paper, we describe an extension to the TDT useful for examining genetic heterogeneity. This extension uses contingency table analyses such as log-linear analysis to test for differences in linkage disequilibrium across levels of one or more moderator variables. We applied these analyses to test for linkage disequilibrium between the dopamine transporter gene (DAT1) and bipolar disorder, as well as for genetic heterogeneity due to sex, diagnostic breadth, and study site. Using data from two studies (the UCSD/UBC and Cardiff data sets), we found evidence suggesting linkage disequilibrium between DAT1 and bipolar disorder, as well as heterogeneity due to diagnostic breadth and study site.