UC Berkeley

The battleground state is ubiquitous in the discourse and scholarship surrounding American presidential campaigns, but the concept remains poorly understood, measured, and operationalized. The nature of presidential geographic targeting carries potentially significant consequences for the nation's issue agenda, political institutions, and voter behavior, and this dissertation details the need to re-conceptualize the battleground state as both an explanatory and dependent variable if these consequences are to be better understood.

Beginning with the 1960 presidential election, I use an original archival data set collected at the nation's presidential libraries to confront the myths that exist in both popular coverage and much of the existing political science literature about battleground states, and I work to correct the record. Media content analysis establishes the significant increase in attention paid to battleground strategies over time in the press. A conceptual analysis highlights both the stability and the evolution of the battleground state concept. The archival records reveal the presidential campaigns' multiple goals, the various geographic strategies adopted to meet those goals, and the elements campaigns use to prioritize geographic areas. The findings challenge many leading assumptions and expose misconceptions made about battleground strategies, and I suggest ways to improve our understanding of the concept. Next, a constitutive analysis using multiplicative interaction models explores the preconditions that guide campaign classification and campaign resource allocation patterns in presidential elections. This analysis demonstrates a fundamental shift in the factors that predict state battleground status and offers more evidence of why it is necessary to more rigorously conceptualize the battleground state. Finally, a causal analysis of the effects of the battleground state on voter evaluations of the candidates reveals the critical link between conceptualization and measurement validity. I demonstrate that different levels of measurement tell us very different stories about the causal processes of campaign effects, and I argue for the increased use of a categorical dummy variable to measure battleground status. Finally, using a block recursive model, I demonstrate that the inclusion of multiple campaign mechanisms and campaign classifications of the states in battleground effects models clarifies the direct effects of different strategies on voter behavior. In presenting these and other findings, I improve our understanding of the battleground state concept and enhance its usefulness as a tool for future research.