Recently in the Atlantic, Yascha Mounk highlighted new polling data showing that Democrats and Republicans believe their counterparts from the other party hold more-extreme views than they actually do. The study he cites titles this distorted view “The Perception Gap,” a finding consistent with some academic studies but not others.
Mounk makes an increasingly common assumption about partisanship: “What is corroding American politics is, specifically, negative partisanship. … America’s political divisions are driven by hatred of an out-group rather than love of the in-group.” He argues that this hatred grows from misunderstanding one another’s views.
But is it true that modern American partisan identity is predominantly negative?
Most partisans do see the opposing party very negatively
Many observers now regularly argue that, as Thomas Edsall of the New York Times puts it, what motivates partisan voters is loathing rather than loyalty. For decades, many have treated elections as a choice between the lesser of two evils. And, to be sure, a great deal of outstanding scholarship has focused on the negative and emotionally charged nature of modern party identity. Many Democrats and Republicans say they are unhappy with both parties.
What’s more, partisans now have more negative feelings about the other side than they did just a few decades ago. Members of a party react more strongly against positions supported by the other party’s leaders than they react in favor of positions their own party’s leaders support. Rank-and-file Democrats and Republicans are increasingly willing to dehumanize members of the other party. There are even hints of tolerance for partisan violence. So there is little doubt that negative attitudes are a central feature of modern polarization.
Wanting to know more, Carlee Beth Hawkins and I set out to study American party identity in light of new research in social psychology. What we found was surprising, given all the research and discussion about negative partisanship. Most American partisans do not appear to have a primarily negative party identity. And when they do, it seems to have more to do with how they feel about their own party than how they feel about the other party.
Here’s how we did our research
New social science research proposes two kinds of identities: “negational” and “affirmational.” In affirmational (or positive) identities, individuals define themselves by belonging to a certain group. In negational (or negative) identities, by contrast, they define themselves by not belonging to a group.
To measure whether people’s party identities were more negative or positive, we adapted a simple survey item developed by social psychologists: “Sometimes we define our identities in terms of the groups we belong to and other times by the groups that we do not belong to. When it comes to political parties, would you say your party affiliation is based more on belonging to your own party or NOT belonging to the other party?” Possible responses ranged from “strongly based on identifying with my own party” to “strongly based on not identifying with the other party,” with an “equally” option in the middle.
We have posed this question to respondents in many surveys, including large, nationally representative samples reached as part of the Cooperative Congressional Election Study, a major survey fielded online by YouGov.
To ensure that respondents weren’t simply unwilling to admit a negative party identity, we also used a set of implicit association tests. These tests use the speed with which respondents are able to sort images and words to measure deep, pre-introspection associations. The results from these tests were consistent with the results from the more direct survey questions.
Are Americans more likely to identify with their party — or against the other party?
In each study, we find the same thing: Partisans largely identify with their own party rather than against the opposition. Let’s look, for example, at respondents in the Cooperative Congressional Election Study that was run during October and November 2016, a historical moment that’s often discussed as a period of particularly negative partisanship. Only 17 percent of partisans said their party identity was primarily negative. Fifty-six percent reported they positively identified with their party, and another 27 percent described their partisanship as equally negative and positive. As the figures below show, both Democrats and Republicans claimed a predominantly positive party identity.
So what kinds of people identify with — or against — a party? Among both Democrats and Republicans, we found a strong relationship between how intensely partisan respondents say they are and how positive or negative they say their party identity is. “Strong” partisans claim more-positive party identities than “not so strong” partisans. Only leaning Democrats and Republicans (those who initially told us they were independents but then acknowledged being “closer” to one party or the other in a follow-up question) had a negative party identity, on average.
Americans don’t like the opposing party, but they mostly still identify with their own.
More and more, partisans report feeling negatively about the opposing party. We often measure such feelings using survey items called feeling thermometers, in which people rate the parties on a “thermometer” scale ranging from 0 (very cold) to 100 (very warm). The average rating of one’s own party has remained relatively high over the past few decades. By contrast, the average feeling about the other party has plummeted from nearly 50 in the 1980s to the low 20s in 2016.
Does this mean that more individuals’ partisan identities are oriented against the other party rather than toward their own? No. Those party identities — negative or positive — do not appear to have much to do with how people feel about the other party.
There is, however, a strong relationship between the negativity of one’s party identity and average feelings about one’s own party. Democrats who say they identify negatively rate their feelings about their own party at 67 degrees, while those who identify with their party affirmatively average a toasty 81 degrees. Republicans who identify negatively rate their party at 52 degrees, while those who embrace their own rate it at 69.
Negative party identity appears less pervasive than we might have expected, and it is associated strongly with ambivalence about or disaffection with a person’s own party, not repulsion from the other party. So while many Americans misunderstand, dislike and even demonize the other side, they appear to primarily identify themselves as members of their own political tribe.
Alexander Theodoridis (@AGTheodoridis) is assistant professor of political science at the University of California-Merced.