Published online by Cambridge University Press: 25 January 2021
Political scientists, historians, pundits, and campaign managers have often sought to understand electoral politics by examining intrastate political geography. But what practical or theoretical contribution can political geography make when we have the powerful tool of survey research? We use a geographic statistic to identify regional nodes in four states, for the 1928–36 and 1988–2000 presidential elections. By weighting county-level election returns for their contribution to the total statewide vote for each party, we find that traditional regional characterizations of these states' politics are altered dramatically. We find that the parties typically compete on the same turf, making clear sectional distinctions harder to draw. Furthermore, over time within three of these four states, the Democratic vote has become more geographically concentrated, while the Republican vote has become more geographically dispersed. These findings have implications for the organization of statewide governing coalitions, the cost of party mobilization efforts, and the study of candidate emergence and success.