What Really Happened in the Midterms?
A segment of swing voters decided to back Democratic candidates in many critical races.Caitlin Ochs for The New York TimesMore than eight months later, all the data from the 2022 midterm elections is — finally — final. The two most rigorous reports, from the Pew Research Center and Catalist, are finished.And yet despite all the data, there is a piece of the midterm puzzle that still hasn’t quite been resolved: How exactly did the Democrats manage to nearly sweep every competitive House and Senate race, even though they often fared quite miserably elsewhere?The Catalist report suggested it was the turnout, finding that Democrats won “with electorates in these contests looking more like the 2020 and 2018 electorates than a typical midterm.” Pew also pointed to turnout, but with a different interpretation, writing that Republicans won control of the House “largely on the strength of higher turnout,” and found that disproportionate numbers of Biden voters and Democrats from 2018 stayed home.You might imagine ways to square the two claims, but neither report offers a clear way to reconcile these competing stories. Catalist, a Democratic data firm, doesn’t mention a word on the partisan makeup of the electorate, despite possessing the data to do so. The Pew report, meanwhile, is framed around explaining how Republicans won the House popular vote by three points — an important outcome, but one overshadowed by the Democratic hold in the Senate and the razor-thin Republican House majority.Fortunately, our data at The New York Times can help piece together what remains of the puzzle. Over the last few years via Times/Siena College polls, we’ve interviewed tens of thousands of voters nationwide and in the crucial battleground states and districts. This data can be linked to voter registration files — the backbone of both the Catalist and Pew reports — that show exactly who voted and who did not (though not whom they voted for, of course), including in the states and districts that decided the midterm election.The findings suggest that the turnout was mostly typical of a midterm election and helped Republicans nationwide, but there are good reasons to doubt whether it was as helpful to the party out of power as it had been in previous midterms.It certainly wasn’t enough to overcome what truly distinguished the 2022 midterm election: the critical sliver of voters who were repelled by specific Republican nominees, Donald J. Trump’s “stop the steal” movement and the Supreme Court’s decision to overturn Roe v. Wade.At a glance, a typical midterm electorateTo some degree, every midterm leans toward the party out of power, and has an older, whiter electorate. Last November was no exception. Just consider these figures on 2022 voters nationwide:73 percent of registered Republicans (defined by whether someone is registered as a Republican or participated in a recent Republican primary) turned out in 2022, compared with 63 percent of registered Democrats. The 10-point turnout advantage meant Republicans narrowly outnumbered Democrats among 2022 voters given that there are about five percentage points more registered Democrats than registered Republicans by this measure.Just 45 percent of Black and 38 percent of Hispanic voters turned out, compared with 58 percent of non-Hispanic whites, according to data from the Census Bureau. The findings are consistent with data from voter registration files and the actual results, as we reported last fall, along with the Pew and Catalist reports, in showing a weak turnout among Black voters.Voters over 65 represented 33 percent of the electorate, according to the L2 data, compared with just 10 percent for those 18 to 29.All of these patterns are consistent with a typical midterm turnout.The size of the Republican registration advantage is almost exactly in line with the available historical data. It also aligns neatly with our pre-election estimates, which you can see for yourself in our final (and highly accurate) Times/Siena polls.And as we reported in December, this basic story holds up in the battleground states as well. Republicans outvoted Democrats everywhere, including in the very states where Democrats excelled.A hidden Democratic turnout advantage?All of this seems to add up to a stark Republican turnout advantage, powered by an older, whiter and more Republican electorate.But perhaps surprisingly, there are reasons to think the actual turnout advantage for Republican candidates might not have been nearly so large as these figures suggest.Just start with the Pew report, which found that Trump voters were four points likelier to turn out than Biden voters, 71 percent to 67 percent. That’s an important advantage, but it’s less than half the size of the 10-point Republican turnout advantage by registration. The Pew figures actually suggest the 2022 midterm electorate backed Joe Biden in 2020, even though registered Republicans outnumbered Democrats.The Times data suggests something similar. According to our estimates, 69.1 percent of Trump voters turned out compared with 66.7 percent of Biden voters — essentially the same as the Pew figures, though edging even closer to parity.These estimates are based on a statistical model that marries Times/Siena polling data and voter records (including someone’s party registration) to predict how registrants voted in the 2020 election. I’ve forced you through that wonky sentence because it means that these estimates are entirely consistent with and inclusive of all of those various Republican-friendly turnout figures offered earlier: Our estimate is that Republicans outvoted Democrats by 10 points but that Trump voters nonetheless outvoted Biden voters by only two points.Looking at the data more carefully, the source of this disparity is mostly among Democrats. The registered Democrats who stayed home in 2022 were disproportionately likely to be those who sometimes vote Republican. The Democrats who turned out, on the other hand, were especially loyal Democrats who voted for Mr. Biden in 2020. This is partly because of education — midterm voters are more highly educated — but the survey data suggests that this Democratic advantage ran a lot deeper.It’s worth being cautious about this finding. The 10-point G.O.P. turnout advantage cited earlier is essentially a fact. The possibility that the practical turnout advantage for Republican candidates might have been only a third of that or less is an estimate based on fallible survey data. It’s also dependent on accurately surveying a group of people — nonvoters — who are very difficult for pollsters to measure.But the Times and Pew data tell a very similar story, despite very different methodologies, and the accurate topline results of the pre-election surveys add additional harmony. The possibility of some kind of hidden underlying Democratic advantage in motivation is also consistent with other data points on 2022, like Democrats’ astonishing success in ultra-low-turnout special elections.Close to parity in the battlegrounds?The 2022 midterm election was not a simple election decided by a national electorate. It was unusually heterogenous, with Republicans enjoying a “red wave” in states like Florida or New York while other states, like Pennsylvania and Michigan, could be argued to have ridden a “blue wave.”As we’ll see, nowhere near all of the difference between these states can be attributed to turnout. But part of the difference was the disparate turnout, with Republicans enjoying a far larger turnout advantage than they did nationwide in states like Florida, while Democrats did better than they did nationwide in states like Pennsylvania. And since our estimates suggest that the Republican turnout advantage nationwide was fairly modest — more modest than the party registration figures suggest — the estimates also show that neither party enjoyed a significant turnout advantage in many battleground states where Democrats turned in above-average performances.In Northern battleground states like Pennsylvania, Michigan, Wisconsin, New Hampshire and Ohio, Biden and Trump voters turned out at nearly identical rates, according to our estimates.In contrast, Trump voters were likelier to turn out than Biden voters by around 10 percentage points or more in states like Florida and New York. In practice, this meant that the Florida electorate most likely voted for Mr. Trump by double digits, even though he carried the state by just three points in 2020.Most states, including the key Sun Belt battlegrounds like Arizona and Georgia, fell in between the Northern battlegrounds and the red-wave states like New York or Florida.A decisive advantage among swing votersThe resilient Democratic turnout in many key Northern battleground states might seem like a key that unlocks what happened in 2022, but it explains less than you might think.According to our estimates, Biden voters only narrowly outnumbered Trump voters in Pennsylvania and Michigan. But Democratic candidates for Senate and governor won in landslides that greatly exceeded Mr. Biden’s margin of victory. Similarly, Trump voters outnumbered Biden voters in Arizona, Georgia and Nevada, where Democrats posted crucial wins that assured control of the Senate.Ultimately, the Democratic performance depended on something that went far beyond turnout: A segment of swing voters decided to back Democratic candidates in many critical races.For all the talk about turnout, this is what distinguished the 2022 midterms from any other in recent memory. Looking back over 15 years, the party out of power has typically won independent voters by an average margin of 14 points, as a crucial segment of voters either has soured on the president or has acted as a check against the excesses of the party in power.This did not happen in 2022. Every major study — the exit polls, the AP/VoteCast study, the Pew study published this week — showed Democrats narrowly won self-identified independent voters, despite an unfavorable national political environment and an older, whiter group of independent voters. A post-election analysis of Times/Siena surveys adjusted to match the final vote count and the validated electorate show the same thing. It took the Democratic resilience among swing voters together with the Democratic resilience in turnout, especially in the Northern battlegrounds, to nearly allow Democrats to hold the U.S. House.In many crucial states, Democratic candidates for Senate and governor often outright excelled among swing voters, plainly winning over a sliver of voters who probably backed Mr. Trump for president in 2020 and certainly supported Republican candidates for U.S. House in 2022. This was most pronounced in the states where Republicans nominated stop-the-steal candidates or where the abortion issue was prominent, like Michigan.Democratic strength among swing voters in key states allowed the party to overcome an important turnout disadvantage in states like Georgia, Arizona and Nevada. That strength turned Pennsylvania and Michigan into landslides. And it ensured that the 2022 midterm election would not go down as an easy Republican victory, despite their takeover of the House, but would instead seem like a setback for conservatives. More