Will the 2024 election polls be correct this time? “Confusing” signs

Polls for the 2024 presidential election show an excruciatingly tight race between Kamala Harris And Donald Trumpwhich could swing either way, with little to separate the two candidates across the seven battleground states that will likely determine the outcome.

But since 2016, when republican Candidate Trump wins surprise victory over Democratic nominee Hillary Clintona result that was not widely or explicitly expected in national polls, the polling industry has come under scrutiny for its accuracy.

Similar problems in later years, such as overestimation of presidential support. Joe Biden Trump’s outperformance in the 2020 race and Democratic under-voting in some key 2022 midterm races keep the issue alive.

But polling companies have done a lot of work to revise and refine their models, trying to recalibrate their assumptions to better capture a representative sample of American voters and thus reestablish their statues as barometers of the nation’s thoughts and feelings.

And this is the key point: polls should measure temperatures in the moment, a snapshot in time, not an exact forecast of what will happen later. Events happen, minds change, and so the results are different from what the polls implied.

On November 1, 538, a company that takes a weighted average of polls and models the results to give a probability-based forecast of who will win, reported that its simulations show Trump winning 53 out of 100 times to Harris’ 47. i-neck.

However, perhaps this is all illusory. There is a possibility that the winner will emerge from the game with a much stronger-than-expected scenario, which some observers say is quite likely. And this will be a result that few sociologists in the country predicted.

With this in mind, Newsweek ask poll experts across the US: Are we on track for another near miss in the 2024 election? And where are polls most at risk of miscalculating support for a particular candidate? That’s what they said.

Charles Franklin, director of Marquette Law School

Next week we’ll find out whether the polls were generally correct this year or not. Clearly, polls and averages point to a very close race in swing states and the national popular vote. This contrasts with polls in 2016 and 2020, which showed larger Democratic margins. So, at least this year, the polls show that we really don’t know who will win because it’s so close. We shouldn’t be surprised if Harris wins, and we shouldn’t be surprised if Trump wins. Both seem equally likely to win.

Mike Traugott, Research Professor Emeritus, Institute for Social Research, University of Michigan

Support for Kamala Harris and Donald Trump is largely entrenched, with very few people left to persuade. So the outcome will depend on turnout and each side’s ability to get supporters to the polls or get a ballot in hand.

This means that the accuracy of polls will depend on their turnout patterns—how they estimate likely voters.

Moreover, I expect to see a very large gender gap in support for the two candidates, perhaps enough to sway the outcome in favor of Harris. For example, if you look at national polls, although most show the races to be very close, their internal data shows gender differences of varying sizes.

This is puzzling in many ways: Women make up the majority of voters and vote at higher rates than men, so this could explain any discrepancies between the final poll estimates (Sunday and Monday) and the actual voting results.

John Zogby, Senior Partner, John Zogby Strategies

I firmly believe that polling in battleground states was strong in 2016. They detected Clinton’s declining support trajectory, enough to indicate she could lose.

If we only look at the polls the day before and treat them as “forecasts,” then it is the experts who will fail, not the pollsters. I think the polls show it’s a tie at the moment.

By the weekend we should have time to find out whether the devil will break in one direction or another. Think of surveys as a series of photographs, not as predictions.

Josh Clinton, co-director of the Vanderbilt Poll

Who knows. Equally likely, I think the polls will do better because this could be a repeat of 2020, and pollsters responded to the polling misses in 2020 to change the way they adjusted their results (which is why we see so many connections in polls); that polls will underestimate Trump even with adjustments if new voters support Trump and don’t take part in the polls (as they did in 2020), or perhaps support for Harris will be underestimated because pollsters are adjusting to eliminate the Democratic/Harris lead into the raw data because they believe Harris’ leadership is an issue with who will respond, not who the electorate will be.

Christopher Wlezien, Hogg Professor of Government, University of Texas at Austin

As for whether we’re on track for another poll miss, it’s important to keep in mind that polls are often wrong: the presidential two-party lead nationally is by an average of 2 to 3 points late in the campaign. .

As we move further away from the election, the numbers are getting higher, and that’s not surprising. This is important to keep in mind when evaluating polls, especially where the margins are small and performance is assessed based on correctly identifying the winner, i.e. polls may get close to the final percentage of votes but still point to the wrong winner.

And then there’s the Electoral College, which looks at polls in states where the errors tend to be larger than what we see nationally.

Now, if you’re wondering whether we’ll see above-average errors this time around, it’s certainly possible, although I suspect the pollsters will do better than they did in 2020 because they have incentive and have put different things into it changes. end.

How much better I don’t know, since it’s hard to say what exactly they do and to what effect, given that it’s not easy to know who will be casting votes, even with significant early voting, and to get them to respond. to surveys.

We don’t know that the decisions pollsters make in making their estimates (can) make a big difference to the polling results they report.

Will they better represent Trump voters who were seemingly overlooked in 2020? Were they, as some analysts suggest, over-corrected? What about new voters? All this remains to be seen.

Meanwhile, given history, don’t be surprised if one of the candidates underperforms or overperforms in the polls in some meaningful way.

Courtney Kennedy, Vice President of Methods and Innovation, Pew Research Center

There are reasons to be optimistic about the accuracy of this year’s polls, but there are also reasons to be pessimistic.

If there is a survey miss, it will not be for lack of effort. My colleagues and I conducted a study that found that the majority of the nation’s pollsters (61 percent) have changed their methods since 2016. That is, they changed either the way they sampled people, or the way they interviewed them, or both.

Compared to 2016, today’s surveys are more likely to provide people with multiple ways to participate (for example, online or by phone). This may help to reach a more representative group of people, as no one approach will suit everyone.

Additionally, the 2022 midterm election results were generally accurate, despite a wave of partisan polls predicting a wide Republican victory. FiveThirtyEight found that “polls in 2022 have been more accurate than in any cycle since 1998, with virtually no bias in favor of either party.”

But there are also reasons to be very cautious about pre-election polls this year. There’s no denying that both times Trump ran for president, most polls underestimated his support.

The main reason seems to be that Trump supporters are less likely to respond to polls than people who are otherwise similar to them (e.g. people of the same age, education, race, etc.). This pattern is not easy for sociologists to identify.

There are some things pollsters can do that we know can help (e.g., give people the ability to take the survey offline, make sure there is an adjustment in the weighing protocol to get a share republicans against Democrats correct).

But this tactic may not be completely effective. Until the votes are counted, we won’t know whether the changes pollsters made were entirely effective.

In recent cycles, the risk of polling error has been particularly high in states like Wisconsin. One reason is that some pollsters pull data from state voter files. In Wisconsin, the data that is added to the voter file—the data that tells pollsters which voters are Republicans and which voters are Democrats—is riddled with errors.

Recent Study found that “less than half of the (registered voters) with a given party affiliation in Wisconsin identified with the imputed party affiliation, and only a third of those identified as Republicans on their voter files identified themselves as Republicans when interviewed.”

In fact, nearly 50 percent of those identified as Republicans on voter rolls report voting for President Biden, much higher than support among likely Republicans in other states. Somewhat similar patterns are observed in Minnesota and Michigan.”

When it comes to demographics, it is especially difficult for sociologists to get accurate information about young people. One reason is simple: sample sizes of young people tend to be small (eg, 100 to 200 interviews), which means the estimates will be quite noisy.

Additionally, young people are more likely to move locations and change phone numbers, making it difficult to obtain current contact information for them.