Describes the eight geographical districts into which St. Joseph County was divided for analysis of the effects of an experimental housing allowance program conducted by the Housing Assistance Supply Experiment. The author uses an analysis of variance model on five demographic and twelve housing characteristic variables to test the appropriateness of the district boundaries: If the districts are well chosen, a statistically significant amount of variation in each variable will be explained by the districts. The results show that for the five variables, the districts do explain a small but statistically significant amount of variation in each variable. For the twelve variables, the districts explain a larger amount of variation. Conclusions are that the districts are well chosen.