Ask Americans a single question — do you think of yourself as a Democrat, a Republican, or an independent? — and you can watch a political universe rearrange itself. The American National Election Studies has asked exactly that, in nearly the same words, since the Truman administration. Pooled and weighted, its seventy years of answers do not show a country that simply drifted left or right. They show a country that re-sorted: the same voters, redistributed across the same two parties along entirely new lines.
Three of those lines do most of the work. The old class divide built by the New Deal has inverted. A new divide over education has opened and then widened into a chasm. And a cleavage over religion, almost invisible in mid-century, has hardened into one of the most reliable predictors of the vote. None of these is a story of one party winning. Each is a story of who ended up where.
A near-tie, seventy years in the making
The Democratic plurality of the 1950s has eroded almost to parity. The drama is less in the levels than in the convergence.
Party identification of all American adults, 1952–2024
Weighted share calling themselves Democrat, Republican, or independent.
Background shading marks the party of the sitting president — blue for Democratic, red for Republican administrations. Three-category party identification; leaners are counted with the party they lean toward. Excludes the small share giving no answer.
In 1952 the Democrats held a 23-point edge in party identification — 59% to 36%. By 2024 that lead had collapsed to a rounding error: 48% to 45%. The Republican line did not so much surge as climb steadily, decade after decade, until the two parties met. The headline number of American politics is no longer a Democratic majority; it is a tie.
The country did not move. Its coalitions did.
That flat aggregate hides everything interesting. A party can hold a steady share while completely swapping the people inside it — trading one bloc of voters for another of roughly equal size. That is precisely what happened. To see it, you have to stop asking how many Democrats there are and start asking which Americans they are.
Income used to point left. Now it points right — or nowhere.
The New Deal coalition rested on a simple premise: the working class voted Democratic, the well-off voted Republican. Among white Americans, that premise has quietly reversed.
Democratic identification by income, lower third vs. upper third
Among white Americans, the income groups have swapped their partisan rank order.
Income terciles within each survey year. Restricting to white respondents isolates the class story from the distinct, and very stable, partisanship of voters of color.
In the mid-century, lower-income whites were the more Democratic group — the legacy of the New Deal and a then-solidly Democratic white South. The gap was never enormous, but its sign was unmistakable: less money, more Democratic.
By 2024 the sign has flipped. Higher-income white Americans are now 13 points more likely to call themselves Democrats than lower-income whites — 48% against 35%. The party of the working man, at least among white voters, now does better with the affluent. Income has not become irrelevant to party; it has switched which side it favors.
What displaced it is the subject of the next section — and it is the single most important rearrangement in modern American politics.
The realignment that ran through the classroom
No variable has reorganized American partisanship as completely as the four-year degree. Watch the rank order of education reverse, level by level.
Democratic identification by education, four levels
Each line is one rung of the educational ladder. They begin tangled and end fully separated.
A 44-point swing in the relative partisanship of the most- versus least-educated whites over seven decades — the defining structural change of the period.
For most of the twentieth century, education and income pointed the same way: the well-schooled and the well-paid were Republicans, and they were often the same people. A diploma signaled affluence, and affluence signaled the GOP. Through the 1950s, college-educated whites were more Republican than those who never finished high school.
A college degree was once a Republican credential. It is now a Democratic one.
Then the lines began to converge, slowly through the 1980s and 1990s, and then with startling speed. Somewhere around 2015 they crossed. Today the relationship is not merely reversed but stretched to an extreme: white college graduates split 56–42 Democratic, while whites without a degree split 28–64 Republican — a gulf of more than fifty points. The classroom, not the paycheck, has become the dividing line of white American politics. Scholars call it the diploma divide, and it is the engine underneath the flat aggregate in Figure 1.
A cleavage that did not exist, then did
How often you attend religious services said little about your politics in the 1970s. It now says a great deal.
Republican identification by religious attendance
Frequent and infrequent attenders pull apart from the 1980s onward.
Attendance is unavailable in the earliest ANES waves; the series begins where the question is consistently asked.
In the early 1970s, churchgoing barely sorted Americans by party — frequent attenders were only a few points more Republican than everyone else. The denominational politics of the era ran on different fault lines entirely. Over the following decades a religiosity gap opened and widened: by 2024 those who attend services often are 18 points more Republican than those who rarely do. Worship frequency went from a footnote to a fault line — the cultural counterpart to the educational realignment, and frequently its mirror image.
Cut the data your own way
Pick a demographic to break the country apart by, then filter by race and region. Each panel shows the Democratic and Republican share within that group, every survey year.