Skip to main content
ABC News
Americans Are Shifting The Rest Of Their Identity To Match Their Politics

Welcome to Secret Identity, our regular column on identity and its role in politics and policy.

We generally think of a person’s race or religion as being fixed — and that those parts of identity (being black, say, or evangelical Christian) drive political views. Most African-Americans vote Democratic. Most evangelical Christians vote Republican. But New York University political scientist Patrick Egan has written a new paper showing evidence that identity and politics operate in the opposite direction too — people shift the non-political parts of their identity, including ethnicity and religion, to align better with being a Democrat or a Republican.

Egan used public opinion data collected through the General Social Survey, one of the most reliable measures of Americans’ views of political and social attitudes that we have. The GSS is conducted every two years and surveys a rotating panel of respondents. Some respondents agree to follow-up interviews two years and four years after their initial interview. Egan’s data set was made up of about 3,900 people who were interviewed three times for the GSS surveys, starting either in 2006, 2008 or 2010 (so the most recent data was from people interviewed in 2010, 2012 and 2014). All three times, respondents were asked to rank themselves on a seven-point ideological scale (from “extremely liberal,” to “moderate, middle of the road,” to “extremely conservative.”) They were also asked questions about aspects of their identity that, at least in theory, are non-ideological — questions like: 1) “From what countries or part of the world did your ancestors come?” and 2) “What is your religious preference? Is it Protestant, Catholic, Jewish, some other religion, or no religion?”

There was more inconsistency among answers to these types of questions than I would have expected. For example, about a quarter of people who identified themselves as born-again Christian in at least one of the three interviews either had not described themselves that way in a previous interview or stopped describing themselves that way in a later interview. Nearly half of respondents who identified themselves as lesbian, gay or bisexual at some point during the three interviews did not identify themselves that way in all three (meaning that some people stopped identifying as LGB, while others started to after not having done so at first).

Egan found that these shifts were correlated with political ideology and partisanship:

  • Liberal Democrats were much more likely than conservative Republicans to start identifying as Latino or saying that their ancestry was African, Asian or Hispanic.
  • Conservative Republicans were much more likely than liberal Democrats to become born-again Christians and to stop identifying as non-religious; liberal Democrats were much more likely than conservative Republicans to leave religion and stop describing themselves as born-again.
  • Conservative Republicans were more likely than liberal Democrats to stop describing themselves as lesbian, gay or bisexual; liberal-leaning Democrats were more likely to start identifying as lesbian, gay or bisexual.

In a recently published book, the University of Pennsylvania’s Michele Margolis makes a case similar to Egan’s, specifically about religion: Her research found, for example, that church attendance by Democrats declined between 2002 and 2004, when then-President Bush and Republicans were emphasizing Bush’s faith and how it connected to his opposition to abortion and gay marriage.

I don’t want to overemphasize the results of these studies. Egan still believes that the primary dynamic in politics and identity is that people change parties to match their other identities. But I think Egan’s analysis is in line with a lot of emerging political science that finds U.S. politics is now a fight about identity and culture (and perhaps it always was). Increasingly, the political party you belong to represents a big part of your identity and is not just a reflection of your political views. It may even be your most important identity.

Asked what he thinks the implications of his research are, Egan said that he shies away from saying whether the results are “good or bad.” “I don’t think one kind of identity (say ethnicity or religion) is necessarily more authentic than another (e.g., ideology or party),” he said in an email to FiveThirtyEight.

He added: “Throughout American history, different kinds of identities have always advanced into the foreground or retreated into the background. These results add to the growing evidence that in our current era, political affiliations and beliefs are increasingly at the core of many Americans’ self-conceptualizations.”

By the numbers

The Pew Research Center last month released a study of the 2016 electorate using voter records that are available only after an election, a method that experts say leads to much more accurate results than we get from Election Day exit polls. What do the Pew numbers show?

On the Democratic side, 60 percent of Hillary Clinton’s voters were non-Hispanic white people. At times, political coverage (including my own articles) overemphasizes the roles that Asians, blacks and Latinos play in the Democratic Party. But just because most nonwhite voters are Democrats, that doesn’t mean most Democrats are nonwhite. In fact, white women are the biggest racial/gender cohort in the Democratic Party, according to the Pew data. And as Harvard scholar Theda Skocpol has written extensively, many of the “resistance” groups that have emerged on the left to protest President Trump are led by college-educated, middle-aged white women. On the other hand, it does seem likely that white men will continue to see their power challenged in the Democratic Party. White men represented fewer than a quarter of Clinton voters in 2016, according to Pew — so they are over-represented among House and Senate Democrats.

On the Republican side, 63 percent of Trump voters were white people (men and women) without college degrees. That means more than a third of Trump voters are not in the bloc that seems to be at the center of every journalistic account of the president’s base. About a quarter of Trump’s backers were white people with college degrees, and about 10 percent were non-whites (with and without degrees). White evangelical Protestants were about a third of Trump supporters, so Trump’s coalition is probably more religiously diverse than press coverage indicates.

If you have ideas for future Secret Identity columns, please reach out to me via email ( or Twitter (@perrybaconjr.)

Perry Bacon Jr. was a senior writer for FiveThirtyEight.