Lynn Vavreck
Lynn Vavreck

Lynn Vavreck is a professor of political science at UCLA and co-author of “The Gamble,” about the 2012 presidential campaign. This column appeared on the Upshot blog on the New York Times website.

People often ask me “who these people are” — those who elected Donald J. Trump or those who voted for Hillary Clinton.

They’ll ask, “What’s the single best description of Trump supporters?” My answer often disappoints them. It’s quite simple: They’re Republicans. When they ask about Clinton supporters, the answer is similar: They’re Democrats.

It seems like a frustrating answer, but for more than six decades, party identification has been shaping the vote. Political scientists have long held that party labels do more than just summarize people’s views on issues and policies. They are expressions of an identity. This trait, like many others, may be learned in the laps of our parents and in our neighborhoods when we are young, the same way we learn about our ethnicities or religions.

And it seems we also want to perpetuate that identity in successive generations. A survey of whom American voters prefer for their children’s marriage partners shows just how powerful party identification has become.

To track the power of party identification on voting, every four years since 1952, the American National Election Study has asked a random sample of Americans which party they identified with and whom they voted for in the most recent presidential election.

In 1952 and 1956, nearly all self-proclaimed Republicans (96 percent) voted for Dwight Eisenhower, while nearly three-quarters of Democrats voted for Adlai Stevenson. Party voting remained the norm for Richard Nixon in 1960, 1968 and 1972, with more than 9 of 10 Republicans choosing him over the Democrat. The Democrats came close to this level of party loyalty in 1964 when Lyndon Johnson faced Barry Goldwater.

Some candidates do worse than others. Senator Goldwater didn’t hold his party together as well (only 73 percent of Republicans were loyal to him in 1964), nor did Hubert Humphrey (77 percent), George McGovern (59 percent) or Jimmy Carter (74 percent in 1980). But by the late 1980s, nearly everyone who identified with a party or even leaned slightly toward a party voted for that party’s candidate in the presidential election.

There have been very few deviations from this pattern over the last two decades. Roughly 90 percent of partisans voted for the candidate from their party in every year since 2000.

For all of its unexpected moments, 2016 looks an awful lot like all the other years: There was no meaningful shift in the pattern of intraparty voting.

The stability of in-party voting suggests that either the parties are getting better at nominating high-quality candidates or membership in the parties has become more homogeneous — so that a typical candidate now gets the support of more partisans because partisans are now more alike than they were 50 years ago. There’s some evidence that both of these things are true, but another way to see the power of partisanship as a social identity is to look at something completely different: marriage.

In 1958, the Gallup Organization asked a random sample of Americans a question about what kind of man people wanted their daughters to marry — a Democrat or a Republican. The question read: If you had a daughter of marriageable age, would you prefer she marry a Democrat or a Republican, all other things being equal?

In the late 1950s, 18 percent of respondents said they would want their daughter to marry a Democrat, 10 percent a Republican, and the remainder, 72 percent, either didn’t answer or said they didn’t care.

I asked a representative sample of people a version of this question (that included “son or daughter”) in the week before the 2016 election and found changes that mimic the changes in the intensity of intraparty voting over these same years. In 2016, more people care about the party of their future in-law than cared in 1958, and there is more desire for same-party marriage than there was in the 1950s.

In 2016, 28 percent of respondents said they wanted their son or daughter to marry a Democrat and 27 percent a Republican, leaving only 45 percent to say they didn’t care.

People who identified with a party had even more intense feelings. In 1958, 33 percent of Democrats wanted their daughters to marry a Democrat, and 25 percent of Republicans wanted their daughters to marry a Republican. But by 2016, 60 percent of Democrats and 63 percent of Republicans felt that way.

People in each party now share more similar views on issues and they are more alike in race and ethnicity. Americans are increasingly surrounded by those who are like-minded — and they seem to prefer to keep it that way for the next generation.