1
Introduction
Despite a longstanding recognition that partisanship is the strongest predictor of
individual voting behavior (Campbell et al. 1960; Bartels 2000), very little attention is
devoted to understanding how partisanship mediates the impact of the campaign on
voting behavior. Although scholars successfully tell a story about how partisan
“activation” occurs in campaigns (Finkel 1993; Berelson et al. 1954), they have yet to
confront the important issue of how this activation occurs. To help fill this gap, this
paper investigates how three aspects of the campaign environment – campaign time,
differences in advertising volume between the campaigns, and the types of appeals made
by the campaigns – impact the voting behavior of partisans.
Using data from both the 2000 Annenberg National Election Study and the
Wisconsin Advertising Project, I find evidence that peripheral partisans – weak
identifiers and independent leaners – respond to variation in the volume and substance of
campaign information flows, although not always in a manner consistent with my
theoretical expectations. Interestingly, the evidence does not support the existence of a
natural activation process. Deviations from theoretical expectations notwithstanding,
finding that partisan activation is influenced by presidential campaigns means that the
impact of campaigns on voting depends heavily on an individual’s partisanship,
particularly by reinvigorating partisan attachments through the party’s core issues.
Previous Research on the Mobilization Model of Campaign Effects
Berelson et al. (1954) were the first to attack the subject of voter mobilization in
detail. Their investigation of campaign dynamics in Elmira, New York found that
processes of social discussion and reinforcement were the foundation of partisan
Convention