A woman gives a thumbs-up to a passing car that honked in approval of pro-Trump demonstrators at Los Angeles International Airport on Feb. 4. (Reed Saxon/Associated Press)

With Donald Trump in the White House, observers are still asking what in his message resonated with enough voters to put him over the top in the electoral college. Theories include economic anxiety, racial resentment, authoritarianism and much more. As the nation debates the new president’s dramatic initiatives to restrict immigration, here is one more possible explanation: a clash of beliefs about what it means to be a real American.

Our research identified four sets of beliefs about what it means to be an American

In our research, we looked at an unusually wide range of attitudes that Americans hold about their nation, as measured by the 1996 and 2004 General Social Survey and a 2012 online panel administered by GfK Research. These nationally representative data from more than 5,000 survey respondents offered us insights into the beliefs of a wide cross-section of the American public.

Using a statistical method called “latent class analysis,” we identified four clusters of respondents, each characterized by a distinct nationalist disposition.

One cluster consisted of Americans we called “ardent nationalists.” These respondents expressed strong national pride, believed the United States to be superior to other countries and defined American identity in ethnically and culturally exclusionary terms. They constituted nearly 20 percent of the sample in 1996 and 2012. In 2004, after the attacks of Sept. 11, 2001, and the wars in Afghanistan and Iraq, 25 percent fell into this group.

An additional 23 to 25 percent of Americans espoused views consistent with what historians have called “the American Creed” of tolerance, inclusivity and pride in democratic institutions.

A third group, the “disengaged,” were largely critical or disengaged in their responses to most of the nationalism questions, a negative mirror image of the ardent nationalists’ views. They made up more than a quarter of the sample in 1996 and 2012, and 18 percent in 2004.

But the largest group, accounting for more than 30 percent of survey respondents in all three years, was the one we call “restrictive nationalists.” Like the ardent nationalists described above, these men and women agreed that someone was “truly American” only if they were Christian, spoke English and had been born in the United States.

But unlike the ardent nationalists, this group expressed a surprising lack of pride in American institutions, including democracy, the arts, economic institutions and science. Moreover, even with statistical controls for political party identification and a host of demographic measures, their attitudes toward immigrants and immigration were significantly more negative than those of other Americans.


Are the “restrictive nationalists” Democrats or Republicans?

Although we have no data on how our respondents voted, Trump’s 2016 campaign rhetoric spoke directly to this combination of attitudes.

Contrary to what many might expect, 45 percent of “strong Democrats” and 46 percent of independents agreed with restrictive nationalist views, compared with 38 percent of Republicans and 17 percent of “strong Republicans.”

Most Republicans were found in two different groups: Those older and less well-educated were mostly ardent nationalists, while higher-income college grads tended to be creedal nationalists.

Demographically, who were the “restrictive nationalists”?

We found proportionally more restrictive nationalists among people without a college degree than among college graduates. Restrictive nationalists had the lowest mean incomes of any group and were especially likely to be evangelical Protestants. They were also more likely to be women than men.

Perhaps most surprising, 68 percent of the African Americans in our sample were restrictive nationalists, as were 55 percent of the English-speaking Hispanics — and 38 percent of white Democrats.

Restrictive nationalism was strongest in the Midwest, which went for Trump in the 2016 election, with almost half of white Midwestern Democrats and more than 60 percent of white Midwestern independents falling into this group.

Compared to the older, more Protestant and more Republican “ardent nationalists,” the restrictive nationalists appeared to be men and women for whom the American Dream had turned sour. Members of this group were significantly less optimistic, self-confident and trusting than were the ardent nationalists. According to other research, these characteristics are associated with mental health and well-being.

Our data were collected long before the 2016 election. But it seems plausible that many white independents and some white Democrats in this group voted for Trump, and that these beliefs may help explain why Democrat Hillary Clinton didn’t do as well among women and voters of color as expected.

To be sure, exit polls suggest that most Democrats and Repub­licans voted for their party’s candidate. But in close races, comparatively few crossovers can make the difference. Even if this group didn’t cross party lines, restrictive nation­al­ism’s resonance among some Democrats and independents may help explain why many of them didn’t turn out to vote. For this group, Trump’s message may not have been as off-putting as pundits had assumed.

What do these different national identities mean for politics?

In their new book, “Democracy for Realists,” political scientists Christopher H. Achen and Larry M. Bartels argue that emotion and identity affect voting more than policy preferences do. We suspect that this contributed to Trump’s victory.

For many U.S. voters, “American” is a central identity — but our research shows that voters hold a wide variety of beliefs about what “American” means. How they understand and feel connected to their nation is likely to influence their response to nativist appeals.

These diverging views can be observed in responses to Trump’s proposed U.S.-Mexico border wall and the executive order banning entry to the United States by nationals from seven predominantly Muslim countries. So far, thousands of protesters and many Democratic elected officials (including almost all U.S. Senate Democrats) have stood squarely in the creedal camp, denouncing the president’s moves as contrary to American values. By contrast, most Republican elected officials, for now, have supported the president’s more restrictive approach or remained silent.

These responses — and the way they reflect different ideas of what it means to be an American — may give us a clue to how the politics of national identity will play out in debates over immigration and other issues in the next few years.

Bart Bonikowski is associate professor of sociology at Harvard University where he studies political culture, with a specific focus on nationalist beliefs and populist political discourse in the United States and Europe. Find him on Twitter @bart.bonikowski

Paul DiMaggio is professor of sociology at New York University who researches culture, organizations, networks and inequality. Find him on Twitter @paul_dimaggio