Little evidence links the strategic decisions of campaigns to individual-level voting behavior. Yet for campaigns to matter in the way that experts argue, exposure to campaigns must also matter so there should be observable differences in the structure of vote choice between battleground and non-battleground states. Combining presidential campaign data with the Senate Election Study, we show that intense campaigning can activate factors like race, ideology, partisanship, and presidential approval. We find that the campaigns affected different variables in 1988 than in 1992, which we hypothesize is the consequence of campaign messages.
McClurg, Scott D. and Holbrook, Thomas M. "Living in a Battleground: Presidential Campaigns and Fundamental Predictors of Vote Choice." (Jan 2009).