A large class of theoretical models posits that voters choose candidates on the basis of issue congruence, but convincing empirical tests of this key claim remain elusive. The most persistent difficulty is obtaining comparable spatial estimates for winning and losing candidates, as well as voters. We address these issues using candidate surveys to characterize the electoral platforms for winners and losers, and large issue batteries in 2008 and 2010 to estimate voter preferences. Questions that were answered by both candidates and citizens allow us to jointly scale these estimates. We find robust evidence that vote choice in congressional elections is strongly associated with spatial proximity. Individual-level and contextual variables commonly associated with congressional voting behavior condition the importance of spatial proximity for vote choice, yet ideological considerations still continue to play a substantial role in vote choice. Our results have important implications for theories of voter decision-making and electoral institutions.