1
High Fidelity?
Presidential Campaign Content and the Lack of Media Accountability
Lynn Vavreck, University of California at Los Angeles
•
Scholars of campaigns have long struggled to operationalize and measure important
independent variables. Recently, data on candidate visits, advertising buys, and the
content of advertisements have improved our ability to answer the fundamental question:
Does the behavior of candidates in campaigns matter to voters in elections? In this paper,
I analyze the advertising and stump speech content of modern presidential campaigns
from 1952 to 2000. I then compare what the candidates said during their campaigns with
the contemporaneous coverage of the candidates in the New York Times. Results show
that candidates use their advertisements to deliver different messages from their speeches.
And while competing candidates mainly talk about different things during their
campaigns, the news coverage of the campaign reduces the day’s events to one
controversial topic. In some cases, this topic is something neither of the candidates
mentioned in their campaigns.
It may seem obvious that if you want to know whether a presidential campaign
had any effects on voters, you ought to know what that campaign was about a priori, so
you know where to look for interesting and important effects. The content of the
campaign message, however, tends to be a missing element from investigations of
campaign effects. Instead, scholars have looked at individual level characteristics of
voters or candidate level treatments and evaluated the effects of variation in candidate
effort or campaign tone on voters (for example Lazarsfeld et al 1948; Campbell et al
1960; Bartels 1986; Johnston et al 1992; Holbrook 1996; Shaw 1999; Geer 1998;
Vavreck et al 2002). The work shows that campaigns can have important effects, and it
is an excellent set of ideas from which to begin thinking about how to incorporate a
measure of campaign content into the search for campaign effects.
•
This paper is part of a larger project called More than Minimal Effects: Presidential Campaigns, National
Context, and Candidate Constraints. Funding for this project has been provided by The Rockefeller Center
at Dartmouth College, Hanover, NH and UCLA’s Faculty Senate Committee on Research. I thank Larry
Bartels, John Geer, Tom Patterson, and Richard Boyd for their comments. Danielle Montgomery, Jennifer
Blake, and Margaret Coblent provided excellent research assistance on the newspaper content analyses.
Page 1 of 43 
Convention