some republicans assume Former President Donald Trump’s historic poll slump spells disaster for the Democratic Party in November. But four centre-right analysts say the accuracy of this assumption is being confounded by new changes in methodology.
These analysts provided some insight into how the numbers will play out, but acknowledged there is no certainty until after Election Day. For them, the accuracy of the 2024 polls boils down to three main factors: demographic overcorrection, response bias, and what they call “recall vote weighting.”
The current battleground cluster is in the former president’s position. equal or better Vice President Kamala Harris’ comments backed Republicans who claim that polling errors will result in Trump winning again. Additionally, some are comparing recent polls to numbers from the same period in 2020 and 2016, which show Trump’s poll numbers are significantly ahead of the previous presidential election.
Nate isn’t that close. https://t.co/oBZxpiWvfR
— Chuck Armstrong (@ChuckArmstron11) October 29, 2024
For example, the average poll on October 25, 2020 had President Joe Biden leading President Trump by 8 percentage points in Michigan. According to to FiveThirtyEight — and Biden only won the state by 2.8%.
On the same day in 2024, Harris led Michigan by 0.7%. With the same level of poll error, Trump would win Michigan by 5.1 percentage points.
However, is it reasonable to expect that there will be a similar level of poll error in 2024, and that Mr. Trump will win the battleground states by more than 5 points?
Trafalgar Group founder Robert Kahei said in an interview with the Caller that he doubts such a dramatic difference, but still expects Trump to outperform the October poll. said.
“Is that possible?” he said. “Yes, but I think the chances are he will only outperform by a couple of points… 2024 will be different, but we see important hidden issues.” [Trump] Vote. We think the states where Mr. Trump is most likely to significantly outperform the polls are probably Arizona, Michigan, and Georgia. ”
Kahali argues that President Trump is likely to outperform polls in these states. That’s because his company found that undecided voters in these states supported the former president. He said polling companies are having a hard time engaging Trump supporters because they fear public backlash and need to “coax” Trump supporters into filling out polls and guarantee their anonymity. added.
There are several reasons to doubt the idea that Trump’s true status is: so Higher than current numbers indicate. Polling organizations have identified and corrected many of the sources of past mistakes. tout It said mid-2022 is the most accurately predicted cycle since at least 1998, with little bias toward either party.
Featured: DecisionDeskHQ Analyst @MichaelPruser This debunks the idea that polls were suddenly “corrected” to “hidden Trump votes” and are therefore now “accurate.”
“I have no confidence in the accuracy of this year’s polls.”
“Response bias is taught and… pic.twitter.com/YPY6cbpr0B
— Eric Daugherty (@EricLDaugh) October 27, 2024
“The reason they were really bad in 2016, in my judgment, was because they underestimated white people who didn’t have a college education,” author and political consultant Ryan Gardusky told the Caller. . “They had the overreaction bias of college-educated white liberals.”
Polling organizations generally did not take into account the education level of voters in 2016, but some data suggests that this data point has not previously divided as sharply along partisan lines. report From the American Association for Public Opinion Research. (Related: Two weeks before Election Day, Trump leads Harris in key battleground states, poll results)
Although companies corrected their education oversight by 2020, polls still significantly underestimated support for Trump that cycle. Gardusky attributed this to “partisan response bias,” or the tendency for certain demographics to respond to polls at higher rates, which could skew predictions for 2024.
“In July, when everyone on the planet was calling out to Joe Biden: quit racing He was going to lose, and even though the polls showed he won against the upperclassmen, he couldn’t beat them. [in 2020]” he said. “That’s how skewed the response is among older liberals, people who live to see Rachel Maddow over and over again.”
Some analysts doubt that the accuracy of 2022 polls necessarily guarantees similarly reliable predictions for 2024.
“It’s about constituency expansion,” Mark Harris, co-founder of political consulting firm ColdSpark, told the Caller. “In presidential election years, there is a huge increase in the number of voters, and it is difficult for people to answer the phone to participate in elections.”
He reiterated concerns about a biased response, noting that white people without a college education make up about half of the population. of pennsylvania It’s “much harder” for voters to vote.
However, adjusting methodologies to improve accuracy is not an insurmountable challenge. Harris argues that private and internal polls are usually fairly reliable. Public pollsters bear the greatest responsibility for past inaccuracies.
Universities and other polling organizations often cannot afford to conduct statewide surveys with large sample sizes, so they must artificially adjust, or “weight,” responses from underrepresented or overrepresented demographics. He explained that this reduces the accuracy of predictions. Pollsters also use online surveys, which he argues are not as “rigorous” as the tactics employed by private companies.
“It’s easy to get [in a public poll]”Oh, we interviewed Republican women 20 times and Trump only got 60% of them…” Well, if you estimate that the women are 40% of the sample, that’s weird. Something is going to happen,” Harris said. . “People who gain a lot of weight are really gambling.”
“It’s just money,” he continued. “I have nothing against them. I understand why they are the way they are. [weight their polls] — Still cutting corners. ”
18 days left until voting day! Here’s how the 2024 presidential election will compare to the same dates in the 2020 and 2016 elections, according to RCP polling averages. pic.twitter.com/hUug00Bryi
— Matt Margolis (@mattmargolis) October 18, 2024
Ongoing problems with response bias and overweighting demographics suggest pollsters still underestimate President Trump’s performance, but Republicans should temper their expectations. . Harris emphasized that response bias will decrease in the weeks leading up to Election Day. He also warned against “recall vote weighting,” in which pollsters ask voters about their past votes and adjust their samples based on that information. This approach, he argues, overlooks the fact that “no two voters are the same.”
Polling organizations are increasingly using this technique to ensure that Trump supporters are adequately included in their samples, with the aim of correcting previous difficulties in fully capturing the former president’s support. We are hiring. According to In the New York Times. Some critics say the current polls could give Republicans too much support.
Mark Mitchell, head of polling at Rasmussen Reports, defended the practice in an interview with The Kohler.
“There are many reasons why Trump is outperforming in the polls in my state,” he said. “I’m leaning heavily toward a recall vote. That way my numbers will be less sensitive to people who are less inclined or who are voting for the first time. And everything we’ve seen so far… , shows that those people are showing up and raking in the rights.”
Mitchell explained that the recall weighting actually reduces President Trump’s overall support in Rasmussen’s poll, noting that without the recall weighting, the former president would have received “4 to 5 in the national popular vote. It will increase the points,” he added.
“I showed it to Trump. [ahead by] “He’s been at 2% in the national popular vote for the past few months,” Mitchell said. “I mean, he’s probably going to conquer all the battlefields, but no one’s talking about that.”
Despite some advances in polling methodology, some indicators remain that pollsters will once again underestimate President Trump’s support, but the ‘disaster’ narrative suggests Although it seems unlikely that the same level of polling wrongs will be corrected in November, the unique dynamics of each election cycle continue. Accurate predictions are difficult.
