Your Political Counterparts Are Not Moral Monsters
Research shows we frequently underestimate the moral character of voters across the aisle
In a 2017 survey, over 40 percent of both Democratic and Republican voters describe the other party as “downright evil.” While dehumanization of the opposition party is prima facie untrue at a literal level, are Republicans or Democrats justified in judging themselves morally superior to their counterparts? How would we even evaluate that?
Social scientists have shown that we actually can assess the accuracy of some of our perceptions of partisan moral superiority — and that these perceptions are indeed truly misguided. Especially for those of us who are most polarized.
These results help confirm that while most people in both parties are far from downright evil, our contempt toward the political opposition is driven partly by downright false beliefs.
In the summer of 2016, as the Clinton-Trump presidential contest was heating up, psychologists Ben Tappin and Ryan McKay recruited over 1,000 subjects to play an economic game in groups comprised of three Democrats and three Republicans.
Each subject faced three choices: (1) an action yielding $5 for themselves and nothing for others in the group (dubbed by the experimenters the “self-interest” action); (2) an action yielding $2.50 for themselves and just their co-partisans; (3) an action yielding $2.50 for all six people in the group (“collective interest”).
Tappin and McKay found that, for members of both parties, the choices were on average nearly identical — and mostly pro-social, even across the partisan divide.
Approximately 58 percent of the subjects from both parties chose collective interest (Option 3), sacrificing their own payoff and maximizing the payoffs for all group members — that is, from both parties — while less than 30 percent of subjects in both parties chose self-interest (Option 1).
A few years later, during the Trump administration, economist Eugen Dimant conducted a related experiment, pairing up subjects to play a game in which each subject was given $10 with the option to contribute any or all of it to “produce” a “public good.” Any dollar contributed to the public good turned into $1.50 to be shared. But since the two players split the public good 50–50, the contributor only received $0.75 from each $1 she contributed. Thus, if subjects were purely self-interested they would contribute nothing at all, and stronger pro-social motives would lead subjects to make larger contributions.
Dimant found that subjects who disliked Trump contributed (on average) just under $5 when told they were paired with Trump supporters, and a bit over $6.50 when paired with fellow Trump opponents. Trump supporters contributed around $5 when paired with opponents, and just over $6 when paired with fellow supporters. Once again, nearly identical behavior was observed across the two groups, and this behavior was substantially pro-social, even when subjects were paired with political opponents.
You might ask: “Well, what does behavior in contrived experiments with small stakes have to do with polarization? Maybe polarized people already know, or wouldn’t be surprised to hear, that members of both parties behave similarly in situations like these?”
Actually, polarized subjects in these experiments typically expected “out-partisans” (subjects on the other side of the aisle) to take morally inferior actions. (Determining what constitutes a moral action is sometimes tricky, but we can infer that a subject likely took a pro-social action for moral reasons when the subject took the action despite it being personally costly.)
In Tappin and McKay’s experiments, Republicans who chose the most pro-social action (collective interest) thought nearly 60 percent of co-partisans would make the same choice. But they also thought just 50 percent of Democrats would do this. Democrats who chose collective interest thought nearly 70 percent of fellow Democrats would make the same choice — yet they only expected slightly less than 40 percent of Republicans to choose collective interest.
Tappin and McKay also asked participants questions about perceptions of moral traits for typical voters in each party, and tried to measure the degree to which subjects felt their own party to be morally superior to the other party. Members of both parties who were most morally polarized (in the top half of their party’s moral polarization distribution) and chose collective interest (indicating that they believed this was the morally best action) underestimated the likelihood that out-partisans would take the same pro-social action by over 30 percentage points.
Dimant found similar results: that subjects who were most polarized had beliefs about the other party’s behavior that were most pessimistic — and most objectively incorrect.
There have been other studies relevant to better understanding our cross-partisan perceptions. Experiments similar to the ones I described above have also found consistent evidence that Democrats and Republicans exhibit about the same level of pro-sociality.
In a study using more intense, though hypothetical stakes, political scientists Michael Barber and Ryan Davis asked survey respondents to consider the “trolley problem,” the classic philosophical thought experiment, but with a partisan twist.
In the trolley problem you are told a train is barreling down a track and bound to run over and kill five people stuck on the main track. You can flip a lever and switch the tracks but then the train kills a different person stuck on a side track.
Barber and Davis found that, without knowing anything about the people stuck on either track, just over 90 percent of respondents from both parties were willing to flip the switch, indicating they thought it was the right thing to do. When told the single person on the side track was a co-partisan and that the five on the main track were out-partisans, willingness to make the sacrifice did decline, but choices were almost identical for the two parties, with around 60 percent of each party flipping the switch to save the five out-partisans. The majority of both parties still did “the right thing” and saved the out-partisans.
Barber and Davis didn’t ask subjects about their perceptions of out-partisans’ choices so I conducted a pre-registered study to assess the accuracy of these perceptions myself. In an incentivized survey conducted on October 20th, I asked members of both parties to guess the percentage of members of each party that would sacrifice a co-partisan to save five out-partisans, with relatively large bonus payments for correct answers.
Democratic respondents guessed that around 58 percent of fellow Democrats would make the sacrifice (to save out-partisans), but that only 48 percent of Republicans would do this (versus a true value of 60 percent). Republican respondents guessed that around 53 percent of fellow Republicans would make the sacrifice, but just 48 percent Democrats would (versus a true value of 58 percent).
So, respondents from both parties underestimated the chance an out-partisan would sacrifice their own to save the opposition by at least 10 percentage points.
I also solicited a standard measure of respondent “affective polarization” (difference in warmness of feelings toward the two parties), and found that, again, the most polarized respondents were once more most prone to under-estimating out-partisans’ moral choices relative to choices made by co-partisans. Of those respondents who said they’d make the sacrifice themselves, the most polarized Republicans guessed that 63 percent of fellow Republicans would make the sacrifice but only 54 percent of Democrats would, and the most polarized Democrats guessed 66 percent of Democrats would make the sacrifice, but just 52 percent of Republicans.
There is consistent evidence that, on both sides of the aisle, those of us who are most polarized underestimate aspects of the moral character of those on the other side.
Yes, the causality driving this result likely runs in both directions: beliefs affect feelings, and feelings affect beliefs. (We dislike them because we underestimate their character, and underestimate their character because we dislike them.) But the logic of the former effect — that beliefs about character drive feelings — is particularly clear intuitively, and supported by relevant literature.
It is worth noting again that the subjects in these experiments, demonstrating similar moral behavior regardless of party, are rank-and-file Democrats and Republicans — not politicians or other elites.
I am certainly not saying that we don’t have deep political problems in this country, nor that political actors in both parties are equally culpable for our problems. I am simply calling attention to the evidence that the voters in both parties are not just fellow humans who share common beliefs and preferences more than we realize — but also make similar, and mostly good, moral choices, by standards we share, more than we know.
In these politically tense times, let’s celebrate and share such good news.
