Scandals and Elections: How Much Does Vitter's Past Matter?
The most fascinating feature of the Louisiana gubernatorial race is how resistant the polls have been to the ‘fundamentals’. For some time now, I have reiterated how these fundamentals – particularly the influence of partisanship – are highly predictive of election outcomes. Yet, in a state that generally favors Republicans in contemporary elections, the polls (as of last week) still show that Democrat State Representative John Bel Edwards leads over Republican U.S. Senator David Vitter. Why is this particular election resisting the fundamentals (so far)?
One common suspicion is that Vitter is held back by his prostitution scandal. While this contention seems reasonable on its face, how much do we really know about how scandals impact elections?
No one at this point can say with 100% certainty exactly how much Vitter’s specific scandal impacts his performance in this particular race. But, it’s worth considering what political science research has revealed about how and when scandals affect elections. Here are some key findings:
1. Scandals Cost Candidates about Five Points on Average
Almost all of the political science research has focused on members of the U.S. Congress. [Nope, I could not find any example of research investigating what happens when a scandalized sitting-U.S. Senator runs for governor in the state of Louisiana, in case you were wondering.] A 2013 study by Scott Basinger of the University of Houston estimated the average effect of scandals on incumbents’ vote share is a five point drop.
So, the effect of scandal (on average) is real but not huge. Nearly three quarters (73%) of scandal-tainted incumbents make it to the general election; the others either decline to seek reelection or lose in the primary. Among those who make it to the general election, 81% win reelection. This reelection rate is 15 percentage points lower than for non-scandalized incumbents, but it still marks a very high rate of electoral survival.
2. Media Coverage Matters, Maybe
Here the evidence in mixed. Media coverage can increase the impact a scandal has on how voters decide. Coverage actually has two effects. First, it boosts public awareness of the scandal. After all, most voters who learn about a candidate's scandal do so from what they see in media coverage. Second, the amount coverage of a scandal impacts how important it becomes in voter decision-making. Scandals tend to have a bigger impact when they are more prominently covered by the press.
But it does not always work that way. In the late 1990s, the media covered Bill Clinton's extramarital affair with Monica Lewinsky extensively. However, studies of public attention and opinion during that period show only about a third of Americans were following the news of the scandal. Clinton's approval ratings remained high.
3. The Impact of Scandals are Constrained by Voters’ Partisanship, Ideology, and Prior Opinions of the Candidate
Part of the reason that scandals do not generally have huge effects is that voters are not blank slates. Voters come to elections with preexisting partisanship, ideology, and opinions of the candidates. Many studies have shown that these pre-existing attitudes can mitigate or magnify the impact of scandals. Generally, the voters who are most responsive to scandal are those least likely to have supported the scandal-tainted candidate in the first place. In other words, we generally are more willing to give a pass for impropriety to candidate with whom we share partisanship, with whom we agree ideologically, and with whom we already like.
4. Sex Scandals Don’t Hurt As Much As Other Scandals
The five point scandal effect represents an average, which varies across specific elections but also across types of scandals. The average effect of sex scandals is also about five points – less than for corruption scandals (about an eight point effect) but more than other political scandals like campaign finance violations (no effect).
A 2011 study by David Doherty of Loyola University-Chicago, Conor M. Dowling of the University of Mississippi, and Michael G. Miller of Barnard College uses a clever experiment to see how voters respond to information about different kinds of scandals. The authors show that a financial scandal like tax evasion has a stronger negative effect on voter support than a scandal involving marital infidelity. The clever part is that the authors also examine how these different scandals impact voters’ evaluations of the tainted candidate’s job performance separately from their evaluations of the candidate as a person. They find that voters make clear distinctions between job and personal evaluations. The marital infidelity scandal had a strong negative effect on personal evaluations of the candidate but almost no effect on evaluations of their performance in office. Thus, it has a smaller impact on votes than the tax evasion scandal, which has a strong negative effect on both job and personal evaluations.
There is an exception, however. When the scandal – whether a sex scandal or otherwise – is somehow linked to an abuse of power it has a stronger negative effect on job evaluations and voter support. The authors offer a pair of anecdotes to illustrate the difference played by an abuse of power:
“For instance, New York Congressman Vito Fossella was involved in a highly visible infidelity scandal in 2008 when it was revealed that he had fathered a child with a woman other than his spouse. However, the affair did not involve his public responsibilities. In contrast, the infidelity scandal that ended John Edwards’ 2008 presidential campaign included allegations of using campaign funds to buy the silence of involved parties.”
5 Social Conservatives Respond Especially Negatively to Sex Scandals.
In their experiment, the researchers also find that social conservatives are especially likely to punish candidates tainted by sex scandal. Among these voters, the impact of sex scandal is not simply confined to personal evaluations of the candidate but extends to overall support as well. [But the study does not indicate whether this impact is any different when the scandal-tainted candidate is thought to be socially conservative as well.]
6. The Passage of Time Reduces Impact
The same trio of authors also examined the lingering impact of scandals in a second study from 2014. The authors found that scandals from two decades earlier had only about half the effect on voters’ evaluations of the tainted candidate as recent scandals.
What Does This Mean For Vitter?
Again, none of these studies speak directly to the particulars of the 2015 Louisiana gubernatorial election. There are several factors that are different: This is not Vitter's first post-scandal election, and this is a gubernatorial election rather than a senate election.
Still, this research provides some guideposts for understanding how Vitter's scandal could impact the election. The amount of time that has passed since Vitter engaged in the behavior underlying the scandal - or, at least, the amount of time since the story first broke in 2007 - may reduce its impact. So, too might the fact that the state generally leans Republican. After all, Vitter all but cruised to reelection in 2010.
On the other hand, if the impact of this specific scandal is equal to the average impact of scandals generally, that may be enough to keep the election closer than where the fundamentals would pull it. To get a sense of this, we need an estimate of how Vitter might have performed in this election if he was scandal-free. Obviously, this is unobservable. But, if we use Cassidy's 56% of the vote in the 2014 senate race as a proxy (which might be generous for Vitter because this election is different in other ways as well), then a five point drop in support due to scandal would just about put Vitter in a dead heat.