This study examines the assertions of urban scholars that the spatial arrangement of urban populations is important in determining the amount of conflict displayed within American cities. The article analyzes the spatial distribution of class groups within 18 cities and the degree of voting solidarity and conflict displayed within segregated and integrated sections of each community. Data were gathered from precinct voting returns for several local referenda in each city to test the following hypotheses: (1) The residential distribution of social-class groups will significantly influence the degree of electoral cohesion these groups display; (2) The spatial distribution of class groups will significantly influence the amount of electoral disagreement between class groups. The study found that communities that displayed segregated class groups had a high degree of class electoral solidarity. Within cities that manifested spatially integrated class groups, however, the electoral cohesion of each class was low. A social-class group located in an area of a city possessing wide class dissimilarity was not likely to vote in agreement with other groups of the same class located elsewhere in the city. The findings of this article suggest that location may be one of the sources of urban political conflicts.