Here’s how we did our research
We wanted to know whether conservatives and liberals reacted differently when sexual misconduct allegations were made about one of “their own.” Drawing on previous research, we hypothesized that conservatives would express less concern about sexual harassment as a problem in society and would be less likely to condemn a member of their own in-group compared with a member of the opposing out-group.
To test this, we partnered with YouGov to conduct an online experiment with about 1,000 Americans. The survey used quotas representative of the U.S. population. We randomly assigned half of the participants to read about either “Bill O’Reilly” or “Harvey Weinstein.” All participants first read the following:
Sexual harassment in the workplace, or unwelcome sexual advances, requests for sexual favors, or conduct of a sexual nature that makes someone feel offended, humiliated or intimidated, is a form of sex discrimination and is illegal in the United States.
Participants were then reminded that either “Bill O’Reilly, a prominent conservative” or “Harvey Weinstein, a prominent liberal” had recently lost his job after sexual harassment allegations. We explicitly mentioned the celebrities’ ideological orientation so that whether or not respondents were familiar with the case, they would still know whether the celebrity was one of “us” or one of “them.”
We then asked participants to make two judgments. First, how big an issue did they believe sexual harassment currently is in the United States? And second, if the allegations are proved true, should O’Reilly or Weinstein go to jail?
Overall, 75 percent of our sample thought that sexual harassment was a somewhat-to-very-serious issue. However, this figure masked strong partisan differences. On a scale of zero (not serious) to 10 (very serious), liberals rated the issue to be more serious (8.3), on average, whereas conservatives rated it slightly above neutral (6.1).
Moreover, out of those respondents who did not think either should go to jail, conservatives made up the majority (49 percent, vs. only 14 percent of liberals). Overall, the odds that a participant would say that Weinstein should go to jail were significantly higher compared with the odds they would say O’Reilly should go to jail. But here, too, we find clear differences by political ideology.
Among liberals in the survey, 90 percent said they thought O’Reilly should go to jail — and about 94 percent said the same about Weinstein. Of course, some respondents may have known that the charges about Weinstein also involved sexual assault, which may have influenced their responses. Nonetheless, in-group favoritism didn’t show up here; few liberals were willing to cut either celebrity a break.
On the other hand, while 90 percent of conservatives said Weinstein should go to jail, just 50 percent said the same about O’Reilly. Conservatives, then, appeared to view the allegations against a conservative media commentator as less serious than those against a liberal movie producer. These results were robust even when adjusting for the influence of gender, age, race, education and whether someone reported having been personally harassed.
Numerous polls show Republicans and Democrats have diverging views about sexual misconduct
Of course, findings from a single experiment do not necessarily settle whether conservatives and liberals view allegations about sexual misconduct differently. Differences in the details of the Weinstein and O’Reilly cases could potentially explain why conservatives showed more in-group favoritism than liberals. But our findings are consistent with much other polling data on the same issue.
For instance, a 2017 YouGov poll revealed that Republicans are much less likely (31 percent) than Democrats (65 percent) to view sexual harassment as a very serious societal issue. Republicans are also less likely than Democrats to want to punish their own when it comes to sexual harassment allegations.
For example, in September, a majority of Republicans (54 percent) told pollsters that Kavanaugh should be confirmed regardless of whether the allegations about him were true. After the sexual harassment allegations against O’Reilly, a 2017 YouGov survey found that only 15 percent of Republicans thought Fox News should have canceled his show.
Meanwhile, Republicans are quick to condemn the other side; 61 percent believed the allegations against Weinstein were credible. But what about Democrats? An even larger proportion of Democrats (69 percent) condemned Weinstein.
Similarly, a 2017 Morning Consult/Politico survey reported that roughly equal numbers of Democrats found the sexual harassment claims against Trump (65 percent) and former president Bill Clinton (64 percent) credible. By contrast, although 69 percent of GOP voters agreed on the allegations against Clinton, only 37 percent believed the allegations against Trump.
Lastly, a majority of Republicans (56 percent) agreed that Al Franken, a Democratic senator from Minnesota who was accused of sexual harassment, should resign. But in the same survey, they were much more forgiving about sexual assault allegations against Roy Moore, a Republican U.S. Senate candidate in Alabama: Only 35 percent of Republicans called for him to drop out.
Conservatives and liberals may rely on different ‘moral foundations’
What explains these differences? One possibility is that liberals and conservatives rely on what scholars refer to as different “moral foundations.”
Moral-foundations theory argues that conservatives and liberals differ in the way they think about “individualizing” values — such as concerns about equality, justice, fairness and harm — and “binding” foundations, which include in-group loyalty, authority and purity. In particular, conservatives are more likely to value strong loyalty to their “tribe,” justify the status quo and resist social change. In contrast, liberals tend to prioritize concerns about social justice, fairness and avoiding harm to others.
In other words, the polarized reactions to sexual misconduct allegations against powerful people may not only be partisan. Rather, they may emerge from more-enduring ideological differences in the way that conservatives and liberals view the world.
Costas Panagopoulos (@professorcostas) is professor of political science and director of big data and quantitative initiatives at Northeastern University, author of “Political Campaigns: Concepts, Context and Consequences” (Oxford University Press, 2017) and editor of American Politics Research.