This is a guest post by political scientist David Peterson of Iowa State University.
The allegations about Herman Cain became public on October 31. Two colleagues and I were fortunate to have a poll of registered Republican and independent voters in Iowa go into the field the very next day, on November 1. We interviewed 1,256 people between November 1 and November 13. This allowed us to track how voters viewed Cain throughout these two weeks.
Did Cain’s support decline? We find that it did: if we model Cain’s support by assuming a linear relationship to the day of interviewing, we estimate that it dropped from 21 points on November 1 to 14 on November 13. However, further analysis suggests that November 7 — when Sharon Bialek held her press conference — was a key date. Indeed, the best-fitting model of support for Cain simply assumes that his support changed after November 7. Prior to that date, the probability that a voter supports Cain is 0.20 (95% confidence interval: 0.18, 0.23). After then, the probability that a person in our sample voted for Cain is 0.15 (95% confidence interval: 0.11, 0.19).
So, our data suggest that the allegations hurt Cain, and that the press conference with Sharon Bialek may have been the turning point. This is the conventional wisdom.
But why do we see these shifts, and which voters are moving? We have a lot of additional questions in our data that help us try to answer this question. In particular, we asked respondents about three character traits that, given lots of research on the topic, should be strong predictors of supporting the candidates. Specifically, we asked how well “honest,” “intelligent,” and “provides strong leadership” describes Cain (as well as Romney, Perry, and Bachmann). Each of these trait perceptions is a strong positive predictor of supporting Cain. One of the possible reasons why Cain is suffering at the polls, then, is that voters see him as less honest, less intelligent, or as a weaker leader as a result of the allegations.
At first glance, this didn’t seem to happen. These questions are 4-scales, with higher values being more positive. Figure 1 shows the three trait ratings, split by if the survey occurred before or after November 7. There clearly is no difference.
One explanation may be which media voters consume. We asked a battery of questions about how often the respondent gets political news from (among other sources) network news and cable news. We don’t know which cable news source these voters are tuning into, but since most of them are Republicans, most are probably watching Fox News. Here are the trends for perceptions of Cain’s honesty indicator, broken down by this measure of news consumption. (Perceptions of his intelligence and leadership show the same pattern.)
The trends over time depend a great deal on whether voters tend to watch network news or cable news. Those who don’t follow either network or cable news do not change their perceptions. Those who watch network news regularly evaluate Cain more negatively after November 7th then they did before November 7th. This is true even after accounting for the fact that they had more negative assessments of Cain before November 7th. In contrast, those who watched cable news regularly viewed Cain more positively after November 7th.
Thus, the effect of the scandal on perceptions of Cain depends on where people are getting their information. Those who tune in to the major networks react as one might expect: they view him more negatively. Cable news watchers, in contrast, report more positive assessments, suggesting that they are rallying behind Cain.
What is the lesson here? Cain has survived the sexual harassment scandal as well as he has because although following the network news does appear to lead people to view him more negatively, cable news watching has a countervailing effect. In a general election, or in an election with a higher level of participation than the Iowa Caucus is likely to have, this is bad news for the candidate. In the Iowa Caucus, however, where those who are going to participate are more attentive to cable news than the average voter, cable news coverage of the scandal may limit the damage to Cain.
