Polls were wildly off in the 2016 and 2020 presidential elections, but pollsters believe the 2024 survey will finally be accurate, according to several people who spoke to the Daily Caller News Foundation. He says he’s not sure if that’s the case.
Polls in 2016 and 2020 generally failed to fully account for non-college-educated white voters, who form a key pillar of former President Donald Trump’s support base. Most pollsters are trying not to underestimate hard-to-reach Trump supporters in 2024, but there are limitations to polling methods for survey groups like young people, who ultimately vote. Adequately predicting the composition of the electorate remains a difficult challenge. Multiple pollsters told DCNF that they are Americans without a college education.
“The single biggest mistake was undercounting non-college voters. In the past, when we undercounted non-college voters, we sometimes unintentionally counted Republicans. The number of non-college voters was underestimated because voters tended to be college-educated and more affluent.College voters essentially over-sampled Republicans and under-sampled Democrats. That means they’re doing it,” Republican pollster John McHenry told DCNF. “This is in some ways a reversal, in part because many of the polls in 2016 and 2020 underestimated non-college voters. Poor quality media opinion conducted online. Surveys show that there are still more voters with a college degree than voters with a college degree.” (Related: It’s not that hard to see why Trump is so popular among men)
Customers watch as former President Donald Trump and Vice President Kamala Harris debate for the first time during the presidential campaign at Bar Tabac on September 10, 2024 in New York City. (Photo by Robert Nickelsburg/Getty Images)
Larry Sabato, director of the Center for Politics at the University of Virginia, suggested in an appearance on CNN in early October that he was unsure whether polling organizations were accurately gauging support for President Trump.
Ultimately, polls predicting how things will pan out in 2024 (which are typically of lower quality than private in-house polls commissioned by campaigns) will need to reach this voter base and properly Evaluation remains a major challenge. Currently, RealClearPolitics polling averages combine many national polls into a single metric. show Nationally, Vice President Kamala Harris has a 0.6 percentage point lead, and President Trump maintains a relatively narrow lead in all seven battleground states.
In the 2016 election, pollsters accurately predicted that former Secretary of State Hillary Clinton would win the national popular vote by a relatively large margin, but with low educational attainment and working-class voters in key states such as Wisconsin and Pennsylvania. They underestimated Trump’s support from whites. Most polls incorrectly predicted that Clinton would win. According to Based on a November 2016 assessment published by Pew Research Center. And in 2020, “it was clear that many, if not most, polls did not fully account for Trump’s strength.” According to And it tells Pew that President Joe Biden’s projected electoral and popular vote victory turns out to be much closer than widely expected. According to For the 2020 cycle of polling ratings by the American Association for Public Opinion Research.
“I don’t believe the campaign is making these mistakes again,” McHenry told DCNF, referring to misleading internal investigations into past campaigns. “If you look at the New York Times pollsters, for example, there are companies in there that I’ve never heard of. That doesn’t mean they’re bad, but it doesn’t mean we’re better. “Should be suspected.” Do they have a good track record? How do they conduct polls? Is it just a quick online panel that suits them and a more robust attempt to make sure people are accurately represented?”
Robert Kahei, senior strategist and pollster at Trafalgar Group, said the “hidden” Trump vote that polling companies have struggled to capture in past elections will still be a reality in 2024. He said that he believed that this was a phenomenon of Trafalgar was one of the most accurate polling organizations in the 2020 presidential race, but “they just happened to ‘call’ some close states wrong, so they didn’t get much recognition.” “but” According to to poll expert Nate Silver.
“I think what we’re seeing is higher hidden Trump votes in 2024 than in 2016. 2024 is the highest, 2016 is the second highest, and 2020 is the lowest hidden vote. ” Kahei told DCNF. “First, it’s hard to get Republicans to respond to surveys. Second, it’s hard to get them to give real answers when you take a survey, because many Republicans are undecided by default. ”
But Democrats have a solid voter mobilization infrastructure in must-win battleground states, and whatever “hidden” Trump votes they have could be answered with higher-than-expected turnout for Harris. says Kahari. (Related: Trump finds the perfect attack as Kamala Harris makes terrible mistakes in battleground states)
WATCH: Kamala gestured and made strange noises while speaking about democracy at City Hall yesterday. pic.twitter.com/anvbTnFCcY
— Daily Caller (@DailyCaller) October 22, 2024
Additionally, Trump campaign pollster John McLaughlin said sampling methods for public polls, and even private polls for low-stakes campaigns, are not as thorough as in-house presidential polls. This is not the case, and it is generally expensive.
“A lot of times, if media polls or campaign polls are doing poorly, they can’t afford to spend money on higher quality polls. So what they’re doing is random calls, and fixed-line calls. We don’t use live interviewers because we do robocalls that are biased towards phone calls and we don’t have enough young people. They use random online databases, but we focus on people taking calls. We want to do that,” McLaughlin told DCNF. “So while sampling is cheap, it’s error-prone, and now that more people are voting early, it’s becoming increasingly difficult to reach voters and ask how they’re voting.”
“The only way to explain the previous bias in 2016 and 2020 is that bias across the media was influencing how the vote was run,” McLaughlin added. “The evidence that the polls are getting better now is that the left is complaining that it’s too good for Trump.”
McLaughlin pointed to October 2020. Washington Post ABC News Poll This is a clear example of the politicization of polls, which show Biden with a 17-point lead over Trump in Wisconsin, where Biden has a 1.1-point margin of victory. It was less than %.
Serinda Lake, an experienced Democratic pollster, discusses the challenges of polling and the factors that can make them less reliable than private in-house polls used for strategic purposes. have their own views.
“I think the most difficult thing about voting is figuring out turnout,” Lake told DCNF. “What will President Trump’s rise be like? Will there be a Harris surge? What will it be?
“It’s not so much the quiet Trump supporters, it’s the Trump supporters that don’t exist in our world. For example, when you look at likely voters, you look at voters who registered in 2020 but didn’t vote. And they’re leaning toward Trump,” Lake continued. “Well, they’re often not included in people’s turnout models. So it’s not that people are silent about their support, and as in past years, President Trump is not included in people’s turnout models. Instead, surge voters don’t exist because they don’t have a voting history.”
Lake told DCNF that many polling organizations run different turnout models to inform their findings, i.e., what portion of the electorate comes to vote and in what numbers. , said it is conducting polls based on different underlying assumptions and evaluating several different scenarios. Another problem with polls, Lake said, is that the media often read too much into the latest numbers.
All content produced by the Daily Caller News Foundation, an independent, nonpartisan news distribution service, is available free of charge to legitimate news publishers with large audiences. All republished articles must include our logo, reporter byline, and DCNF affiliation. If you have any questions about our guidelines or partnering with us, please contact us at licensing@dailycallernewsfoundation.org.