英语论文网

s, we offer a third theoretical possibility: the “differential
effects” hypothesis. While it is too strong to suggest that the differential effects
hypothesis represents a new theory of political advertising, its articulation and
examination is valuable because it has implications that differ sufficiently from
existing frameworks. Fourth, our experimental design includes several features
that attempt to improve external validity.
The article proceeds as follows. Section one outlines the theoretical questions
and issues we address using both existing literature and interviews with the campaign
consultants of Bush and Gore. In addition to sketching briefly the demobilization
and stimulation hypotheses, we also explain the differential effects
hypothesis. Section two describes the experimental research design we use to
examine how political advertisements affect turnout. Section three investigates
the immediate impact of political advertising on viewers’ self-reported likelihood
of voting and evaluates the evidence in favor of each of the three hypotheses.
Section four uses the panel design we employ to analyze the persistence of treatmenteffects on turnout, and section five concludes.
Theory and Hypotheses
Two theories dominate political advertising research—the demobilization andstimulation hypotheses. The names are appropriate as each indicates the expectedimpact of negative advertising on voter turnout. The primary claims of the demobilization
and the stimulation hypotheses address the influence of an advertisement’stone and not its message. To understand the focus on advertising tone,consider three post-WWII trends. First, voter turnout has continued to decline
(Rosenstone and Hansen 1993; Sigelman et al. 1985). Second, political campaigns
have grown in scope, breadth, and cost (Sorauf 1988). Third, campaigns
have become predominantly negative (Jamieson 2001; West 1997).
The demobilization hypothesis argues that a causal relationship exists between
the (negatively correlated) trending variables of voter turnout and campaign negativity
(Ansolabehere and Iyengar 1995; Ansolabehere et al. 1994). Specifically:
70 Joshua D. Clinton and John S. Lapinski

 

Appendix continued
Measures Used in the Analysis
Measure Description
log(Democratic Spending) Advertising spending by Gore and DNC in state after Democratic
convention in dollars (from Kirk Brown of Hickman and
Brown)
Registration rules Number of days before the election a state requires voters to
register in order to vote. Ranges from 0 (same day registration)
to 31 days.
Registered to Vote 1 (“At current address” or “At a different address”); 0 (Else)
Gubernatorial race 1 (Gubernatorial race in state), 0 (no Gubernatorial race)
Senate race 1 (Senate race in state), 0 (no Senate race)
Black 1 (black respondent); 0 (otherwise)
Education 1 (“Less than High School”) to
5 (“Post graduate degree [Masters, professional, doctorate]”)
Attention to Campaign (orig.) 1 (“None”) to 5 (“A great deal”)
Recoded into: Low (1 or 2) otherwise 0; High (4 or 5) otherwise 0
Attention to Campaign (low) Omitted category = 3
Attention to Campaign (high)
Age Age of respondent (must be over 18 and eligible to vote to be a
participant)
Age2 Square of Age
Democrat Indicator variable for identifying as Democrat (1); 0 otherwise
Republican Indicator variable for本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。