Baum: Talking the vote

Baum. 2005. Talking the vote: Why presidential candidates hit the talk show circuit. AJPS.

Overview/Puzzle

In recent years, presidential candidates have taken to appearing on entertainment-talk ("e-talk") shows--the type that generally have very little political content and appeal mostly to those uninvolved in politics. What effects does this have on candidate likeability and on voting?

Arguments/Hypotheses

There are actually 9 hypotheses (see pgs 215 and 217), but for purposes of conciseness, they can be reduced to two main hypotheses:

H1: Content

Compared to news shows, appearances on e-talk shows make fewer references to politics/parties and are less critical of the candidates. Candidates can expect to receive generally favorable treatment.

H2: Effect on voting and on candidate likeability

Low-information viewers, according to Zaller (1992), don't respond to political media because they don't know what it says. But when candidates appear on e-talk, low-information voters get a chance to learn about them. Without this chance, these viewers would probably use only their partisanship to evaluate the candidates. But since the e-talk shows give a positive presentation of the candidate (mainly the candidate's personality), appearances on e-talk will improve perceptions of the other party's candidate and increase the probability of crossover voting (but only among low-information voters). (Since you already like your party's guy, his appearance on TV probably won't change your mind much.)

But as your level of information increases, this effect will disappear, and might even reverse (since you don't like seeing the other guy on TV getting such easy treatment, perhaps).

Data

Testing H1

He codes content analysis from e-talk shows, news interviews, etc. As shown in Table 1, there are far fewer mentions of party, issues, and so on in e-talk shows than in traditional political shows. The differences are significant. As shown in Table 2, candidates get lots of supportive comments but few critical comments on e-talk shows.

Testing H2

Uses ANES data. Table 3 presents a whopper of a regression to show that the set of hypotheses in H2 are correct. Fig 1 shows this clearly. Low-awareness Republicans like Gore more if they watch lots of e-talk, and low-awareness Democrats like Bush more if they watch lots of e-talk. These likeability scores also map onto crossover voting. Among high-awareness partisans, though, the effect is either opposite more mildly in the opposite direction.

Implications

Studies of the effects of media and campaigns have needlessly confined themselves to political media, ignoring the fact that non-political media can also have effects on voters. You don't have to watch the news to get political information.