Sanford's win and Weiner's poll numbers — not to mention Sen. David Vitter's easy reelection in Louisiana in 2010 following a prostitution scandal — are anecdotal evidence that voters are willing to forgive and forget. But there's somewhat more rigorous research to back that up. Indeed, it seems like Sanford's and Spitzer's cooling-off periods of four to five years are pretty much ideal.
A study released in the Social Science Quarterly last month -- by Rodrigo Praino and Vincent Moscardelli at the University of Connecticut and Daniel Stockemer at the University of Ottawa -- looked at every U.S. House race between 1972 and 2006 and compared the outcomes of races where a member had been referred to the House Ethics Committee (88 in total, over that period) to non-scandal races (7,693 overall). As the Pacific Standard's Michael Todd notes, that sampling method doesn't net every scandal. Weiner, for instance, resigned before his case was referred to the Ethics Committee. But it's a pretty good, reasonably objective way of identifying scandals.
It's no surprise that more candidates in scandal races lose, resign or retire. But although their retention rate of 48.9 percent is much lower than the non-scandal retention rate of 86.5 percent, it's still pretty high:
When you control for other factors, like the amount spent on the campaign by each side, the national tilt of that year's election, and the partisanship of the House member's district, having an issue referred to the ethics committee reduces a member of Congress's margin of victory by about 14.5 points, on average. That's nothing to scoff at. But the average reelection margin over the period in question was 31 points, so a scandal tended to reduce that to 16.5 points — still a landslide.
But the really interesting finding had to do with the electoral results years in the future for scandal-afflicted members of Congress:
The solid line is the predicted margin for members of Congress who face an ethics probe, and the dotted line the predicted margin for members who don't but are of comparable seniority. Two years before a scandal, the two barely differ. The year of a scandal, the margins of those facing probes fall, as one would predict. But the two lines converge thereafter, and by four years after the scandal there's barely any difference between thoe candidate who are investigated and those who are not.
Other studies of the House have also found that the electoral cost of scandal is, while real, usually surmountable. Scott Basinger at the University of Houston compiled a database of every scandal between 1973 and 2010, capturing some that the House Ethics Committee methodology of Praino et al would have missed. He found that a scandal reduces a candidate's vote share by about five points, for a reduced margin of 10 points. What's more, the type of scandal matters. Corruption scandals result in an eight-point dip in vote share, sex or financial scandals cause a five-point dip, and fundraising/political scandals don't hurt at all. See Danny Hayes's post on Basinger's work here for more detail. Lisa Brown, then at Cal State - Channel Islands, found that, between 1966 and 2002, having a money-related scandal reduced a House member's vote share in the next election by 6.7 points, and a moral or sex scandal reduced it by 9.2 points. That translates to margin reductions of 13.4 and 18.4 points, respectively.
These studies focus on members of the House and on subsequent House races, so their validity for campaigns for municipal office, such as Weiner's and Spitzer's, is unclear. But there's other evidence to back up the idea that time can improve scandal-afflicted candidates' chances.
Political scientists David Doherty (at Loyola University in Chicago), Conor Dowling (at the University of Mississippi) and Michael Miller (at the University of Illinois - Springfield) ran a national survey that asked respondents to evaluate their likeliness to vote for a state representative candidate who had an extramarital affair with a campaign worker either 20 years ago or one year ago, as well as their likeliness to vote for a candidate who had flip-flopped on abortion either 20 years ago or one year ago. "We find that the effect of either a scandal or a position change (flipflop) that occurred 20 years ago is only half the size of that associated with an identical misstep that occurred one year ago," they concluded.
Other experimental studies have suggested that the personal characteristics of the candidates implicated in a scandal also matter. Adam Berinsky (at MIT), Vincent Hutchings and Nicholas Valentino (Michigan), and Tali Mendelberg and Lee Shaker (Princeton) ran a national survey of white voters in advance of the 2008 Democratic presidential primary in which respondents in the treatment group were presented with articles detailing allegations of extramarital affairs levied against either Barack Obama or North Carolina's Sen. John Edwards. They note that the survey was "conceived and run before (true) allegations of sexual impropriety were raised about Edwards."
They found that the allegations led respondents to rate both candidates as more liberal than a control group did, but the effect on ratings of Obama's ideology was significantly larger than the effect on Edwards. What's more, the effect of the scandal on politically engaged white voters's overall favorability ratings of the candidates was stronger for Obama than for Edwards. "A negative story involving rumors of a sexual infidelity scandal hurts Obama more than it hurts Edwards in both a direct and immediate sense—on his overall favorability rating—as well as indirectly and potentially—through perceptions of his liberal ideology," the authors conclude.
The authors take this as an indication that white voters are more likely to punish black politicians for sex scandals than they are to punish white politicians. White politicians like Weiner, Spitzer and Sanford, then, might be given more of a break for their indiscretions than black candidates in their positions would be.