Skip to main content
ABC News
Stop Using The Midterms To Predict Presidential Elections

Welcome to Pollapalooza, our weekly polling roundup.

The 2022 midterm election hasn’t even happened yet, but speculation about the 2024 presidential race is already underway, with many potential candidates positioning themselves for a possible run — and in the case of former President Donald Trump, doing so quite publicly. Early general election polls are already testing Trump against President Biden, too, with many questioning whether Biden will even run again in 2024, when he’d be 81 years old. Some Democrats have even gone as far as to venture that a rough midterm election for their party — which seems quite possible — might discourage Biden from running again, opening the door for Vice President Kamala Harris or another Democrat to win the party’s nomination.

But while a strong Republican performance in 2022 might influence Biden’s thinking on 2024, it’s a mistake to expect the midterm result to predict how the country will vote in the ensuing presidential election. As the chart below shows, there’s little relationship between the margin in the national popular vote for the U.S. House in a midterm election and the national popular vote margin in the next presidential contest.correlation is a weak -.258 from 1946 to 2020. Of course, there were a handful of presidential cycles with sizable third-party vote shares, such as George Wallace’s 14 percent in 1968 and Ross Perot’s 19 percent in 1992, which complicates calculating the presidential vote margin, but even if you exclude those two cases where a third-party candidate won more than 10 percent, there is still little relationship between midterm and presidential results.

">1 So whatever happens in the House this November, don’t expect it to forecast the results of the next presidential race.

More often than not, midterm and presidential results have been miles apart. For instance, in 1994, the GOP captured Congress in the midterm “Republican Revolution” amidst then-President Bill Clinton’s poor approval ratings, winning the House popular vote by about 7 percentage points. But in 1996, Clinton ended up claiming a sleepy reelection victory by 8.5 points over Bob Dole — a swing of more than 15 points. In fact, on average, there’s been a swing of about 12 points from the midterm to the subsequent presidential contest.

That said, midterm elections and their following presidential races have sometimes run close together: Democrats swept back into power in the 2006 midterm election with an 8-point edge in the House vote, for instance, and then-President Barack Obama won by a similar 7-point margin nationally in 2008. In fact, midterm and presidential elections have run closer together in recent years — but importantly, there is still not much of a relationship despite this. Rather, this seems to be happening because landslide presidential wins are increasingly a thing of the past. Instead, lopsided midterm results for the party not in the White House are increasingly becoming the norm, then followed by highly competitive presidential races.

One big reason midterm and presidential elections have such different outcomes is that they are fundamentally different kinds of contests. Midterm elections, for instance, tend to have much lower turnout than presidential elections, and what’s more, voters from the party that’s not in the White House are often much more likely to turn out. Additionally, some voters switch from the president’s party because of dissatisfaction with the status quo. And because the president’s party usually loses ground in a midterm, you still often see narratives that suggest a president’s reelection chances may now be in peril just after a midterm election. 

Yet presidents can, of course, change their approach following a difficult midterm and/or their opponents in Congress can overstep after winning power, both of which can help improve the president’s public standing. And perhaps most importantly, two years is a lifetime in politics, leaving plenty of time for important factors like the state of the economy or war and peace to change in meaningful ways that will influence the electorate. For instance, a struggling economy, mediocre approval ratings, and a bad midterm can become a roaring economy, “Morning Again In America,” and a landslide victory in that narrow span of time, as was the case for former President Ronald Reagan’s 1984 reelection bid. 

Given all this, Biden and Democrats who are uncertain about the party’s standing probably shouldn’t look to the 2022 election result as a sign of whether the president can win in 2024. It’s true that the midterm outcome could alter the public’s perception of Washington if Biden is facing off with a GOP-controlled Congress. But the actual midterm results won’t really tell us whether Biden is doomed or can win reelection. After all, we have no idea where things will stand in a couple years when it comes to the COVID-19 pandemic, inflation and the overall economy, much less sudden developments like the U.S.’s potential role in the conflict in Ukraine.

Of course, with all this said, it is indeed fun to take a look at the early polling of the 2024 race — trust me, it’s like catnip. Nevertheless, we need to take those early polls with many grains of salt. After all, even polls taken six months out from the general election are only moderately predictive of the final result — and at that point, we usually know who the nominees are.

The 2024 election is likely to be pretty competitive considering how divided our politics are, but the candidates do matter at the margins. So regardless of how many polls test Biden against Trump, time will tell whether we will really see the first general election rematch since 1956 — or if one or both parties end up nominating someone else entirely. But one thing is for certain: The result this November will not guarantee any result for November 2024.

Other polling bites

  • Russia attacked Ukraine early Thursday morning in a conflict that has already killed dozens. But while Americans generally support helping Ukraine, they largely oppose military intervention and think the American government should be minimally involved at this point. According to a Feb. 19-20 poll from Morning Consult, 51 percent of registered voters think the U.S. should impose sanctions, while just 19 percent think the U.S. should deploy troops into eastern Ukraine. And with the caveat that this poll doesn’t account for Russia’s recent attacks, support for sanctions has increased 7 percentage points since Morning Consult last asked this question on Feb. 7.
  • Amidst ongoing efforts to ban books from school libraries, and statewide efforts to restrict how public schools teach race, a Feb. 15-18 poll from CBS News/YouGov found that Americans largely oppose efforts to remove books that contain controversial themes from schools. At least 83 percent said books should never be banned for containing political ideas they disagree with, discussions on race, depictions of slavery, or criticisms of historical figures and events from U.S. history. And over three-quarters of respondents, including 63 percent of Republicans and 89 percent of Democrats, said that public schools should be able to teach about ideas and historical events that might make some students uncomfortable.
  • Fifty-nine percent of American workers who can do their job from home are doing so all or most of the time, according to a Jan. 24-30 survey from the Pew Research Center. That marks a 12-percentage-point decline from those who said the same in October 2020, but the share who say they’re choosing to work from home versus those whose workplaces remain closed or unavailable has almost flipped. Back in October 2020, 64 percent worked at home because the office was closed versus 38 percent now, while 36 percent chose to work from home versus 61 percent now.
  • Healthcare workers still report high job satisfaction, but a majority also say they’ve been struggling with burnout. In a Feb. 9-16 poll from USA Today/Ipsos, 80 percent say they are very or somewhat satisfied with their job. Fifty-nine percent say they feel hopeful or motivated to go to work, while 56 percent say they feel optimistic. But 52 percent report feeling burned out, especially younger workers — 61 percent of healthcare workers under 30 felt this way, versus 45 percent of those 50 and older.
  • Americans are paying less attention to national news, with the decline mostly coming from Democrats, according to a recent Gallup/Knight Foundation report. A poll of 4,221 U.S. adults from Nov. 23-Dec. 3 found that the share who said they pay a great deal of attention to national news (33 percent) was the smallest since the Knight Foundation started tracking that question in April 2018. The biggest drop in attention came from Democrats: Just 34 percent of Democrats said they pay a great deal of attention to national news, versus 69 percent in November 2020. To be sure, though, attention to national news might be down, but it’s still larger than the overall share who say they pay a great deal of attention to local (21 percent) or international news (12 percent).

Biden approval

According to FiveThirtyEight’s presidential approval tracker,2 42.1 percent of Americans approve of the job Biden is doing as president, while 52.5 percent disapprove (a net approval rating of -10.4 points). At this time last week, 41.4 percent approved and 53.0 percent disapproved (a net approval rating of -11.6 points). One month ago, Biden had an approval rating of 41.3 percent and a disapproval rating of 53.5 percent, for a net approval rating of -12.2 points.

Generic ballot

In our average of polls of the generic congressional ballot,3 Republicans currently lead by 2.6 percentage points (45 percent to 42.5 percent). A week ago, Republicans led Democrats by 2.1 points (44.8 percent to 42.6 percent). At this time last month, voters preferred Republicans by 1.6 points (43.8 percent to 42.2 percent).


  1. The correlation is a weak -.258 from 1946 to 2020. Of course, there were a handful of presidential cycles with sizable third-party vote shares, such as George Wallace’s 14 percent in 1968 and Ross Perot’s 19 percent in 1992, which complicates calculating the presidential vote margin, but even if you exclude those two cases where a third-party candidate won more than 10 percent, there is still little relationship between midterm and presidential results.

  2. As of 5 p.m. Eastern on Thursday.

  3. As of 5 p.m. Eastern on Thursday.

Geoffrey Skelley is a senior elections analyst at FiveThirtyEight.

Jean Yi is a former politics intern at FiveThirtyEight.