Clinton and Lapinski: "Targeted" advertising and voter turnout

Clinton and Lapinski. 2004. "Targeted" advertising and voter turnout: An experimental study of the 2000 Presidential election. Journal of Politics 66.

Main Argument

There is a debate in the literature: Do negative advertisements increase ("stimulate") or decrease ("demobilize") turnout? The authors argue that neither is correct; instead, ads target specific issue publics and these issue publics respond to content regardless of tone ("differential effects").

The Lit's Two Arguments

The lit has been divided around two arguments. Since each is backed by evidence from different methods (lab experiments vs voter surveys), it's hard to compare the two.

The Demobilization Hypothesis

Ansolabehere and Iyengar did all sorts of experimental studies that suggested that negative ads increase voter apathy and, therefore, decrease turnout. These effects are most pronounced on independents, since negative ads "Heighten the partisan flavor of political discourse."

The Stimulation Hypothesis

Survey research (mainly ANES) suggests the opposite. Negative ads are at least as informative as positive ads; information increases interest; interest increases turnout; therefore, negative ads promote turnout (see Lau, for example).

A New Argument

Despite the lit's arguments, political consultants seem to think that "what matters most to voters in determining whether or not they vote are the issues and personalitities involved in the race" (p 72). They target their ads at various issue publics, which respond to the issue content of the ad regardless of tone. Since ads have different effects on different subgroups, this is the "differential effects" hypothesis.

Methodology

Overview

An experimental panel survey. Some people are shown positive or negative ads during some of the surveys. Surveys are taken both before and after the 2000 election.

Sample

They used Knowledge Networks' panel. KN gives randomly selected volunteers the opportunity to have free internet and free WebTV if they agree to take a weekly survey.

• [Potential problem: I'm guessing that a lot of people take the free internet, then tell their kids to take all the surveys. That's what I would do. And I'm guessing most kids lie in the surveys.]

Design

• Control group sees no political ads, but takes the surveys (all surveys include both political and non-political questions).
• Treatment groups see some combination of the following: A positive Gore ad, a negative Gore ad, a positive Bush ad, a negative Bush ad.
• Some groups see ads in two separate Waves of the survey.
• Each survey asks people to rate the probability that they will vote in the upcoming election. The post-election survey asks whether you actually voted. A remarkable 82% reported doing so.

Findings

• The authors find no evidence for the demobilization hypothesis. People who saw negative ads were no more likely to report a lower probability of voting (or not voting on election day).
• The authors find weak support for the stimulation hypothesis.
• The authors find the most support for the differential effects hypothesis. For example, Fig 1 shows how age and gender influenced people's responses to a particular ad that targed women and the elderly (prescription drugs). Though the control group showed typical effects (women slightly more likely to vote than men, turnout increases with age), those who saw the ad were different in two ways. (1) Women were much more likely than men to vote, and (2) the elderly were much more likely to turn out, as were the young (who are also effected by prescription drug costs, since they bear the costs). [[]]