15
Discussion and Further Examinations of the PID Latency
Scholars have shown renewed interest in understanding the unique effect
campaigns have on political behavior in recent years. This paper is an attempt to
approach the question of how campaigns matter by examining the cognitive activation of
partisanship attitudes. The models yield strong support for the campaign’s influence on
the PID response latency. Campaigns present a heightened political environment that
allows partisanship to bubble to the top of voters’ minds. Put differently, the campaign
environment can bring individual partisanship to the top of the cognitive funnel,
lessening the time needed to recall the attitude from long-term memory. Response
latencies present an opportunity to tap in to the subtle effects of the campaign while
obviating the bias inherent in many self-reported measures of which scholars are often
dubious.
This paper has thus far focused on the response latency as a dependent variable.
Previous studies have analyzed the role that response latencies play as an explanatory
variable (Fazio and Williams 1986; Bassili 1991; Huckfeldt 1999; Johnson 2004).
Although finding that latencies are sensitive to the campaign is interesting in itself, it is
also appropriate to ask whether the PID response latency influences political behavior as
a right-side-of-the-equation variable. For the remainder of the paper, I put forth a first-
cut analysis of vote defection in the 1996 campaign. The dependent variable of vote
defection is a dichotomous measure of whether a partisan voter defected and voted for a
candidate of another party. For example, Republicans who voted for Clinton or for Ross
Perot are coded as 1. Likewise, Democrats who voted for Dole or Ross Perot are coded
Convention