Why pollsters once again had difficulty measuring the vote for Donald Trump

What goes through the minds of American voters when Donald Trump’s name appears on the ballot? This is the question that pollsters have been asking for eight years and the billionaire’s first victory, which opinion polls could not anticipate in 2016. The day after Donald Trump’s second election to the White House, the question is asked again. emerge, since the Republican’s results exceed the level of voting intention that the polls attributed to him until the day before the elections.

At the national level, the average of surveys calculated by the different media (the New York TimesFiveThirtyEight, The Silver Bulletin) gave Kamala Harris a lead of about a percentage point. But provisional results from Wednesday, November 6, show that it was actually Donald Trump who won the popular vote, with a comfortable margin of around 5 million votes, or a difference of 4 to 4.5 percentage points. This is well above the average national gap since 1988 (2.3 points) and similar to the gap seen in 2020, when the media overestimated the popular vote for Joe Biden by 4 points.

At the level of crucial states, the average of the 2024 polls underestimated Trump’s vote in some of them and to varying degrees. Pennsylvania, where Kamala Harris long held a slim lead of around 1 point, finally saw Donald Trump triumph by a 3-point margin. Nevada shows a similar gap, since the Republican candidate maintains, at the time of writing this report, a 4-point advantage over his Democratic rival, while polls compiled by the media showed a tie between both candidates.

To a lesser extent, the other two Midwestern swing states, Wisconsin and Michigan, show a substantial 2 to 3 point gap between most polls, which gave Harris a lead, and the provisional results, which show Trump winning. with advantage. several hundred thousand votes in advance. But these scenarios are frequently observed: since 2000, polls in swing states typically “miss” the election result by 3.1 points on average. Therefore, specialists relatively expected that 2024 would not be an exception to this rule and that the polls would underestimate one or the other of the candidates by the same order of magnitude. In this sense, the differences observed in the crucial southern states, North Carolina, Georgia and Arizona (1-2 points), can be considered particularly small.

However, all the observed differences point to a slight to moderate underestimation of the vote for Donald J. Trump. Which is not the first time.

Trump voters respond less to polls

This is the third time in a row that pollsters have failed to correctly measure voting intentions for Trump. In 2016 and 2020, opinion polls suffered from the same flaw in all swing states (with the exception of Nevada in 2016).

In 2020, to better represent the Republican candidate’s voters, pollsters took into account the level of education of respondents, a variable that was not “predictive” of Republican or Democratic votes before 2016. In this way they hoped to better represent to the voters. non-graduates, who had voted overwhelmingly for Trump four years earlier. In vain, since the differences were even greater than in 2016, without much impact on public opinion since Joe Biden, who was the favorite, won the elections.

The fact is, pollsters still don’t know for sure why their polls had difficulty correctly anticipating the Trump vote in 2020 and 2016. Especially since during the 2018 and 2022 midterm elections, the polls were relatively accurate.

The specialists’ main hypothesis is that Donald Trump voters would have a lower propensity to answer surveys when asked, which is called non-response bias. The institutes struggle to survey the least politically engaged voters, those who vote infrequently and irregularly. However, the data show that this less politicized electorate overwhelmingly supports the Republican candidate, especially during presidential elections, since they barely vote in the midterm elections, which tend to mobilize the most politicized Americans.

A smaller gap than in 2016 and 2020

And in fact, survey data shows that nonresponse bias exists. In 2020, the New York Times had measured a 20% gap between the response rate of registered Democratic voters and that of Republican voters. Four years later, the New York newspaper’s surveys measured an equivalent differential (16%) among the white electorate, which made the main political analyst of the Timesthe day before the elections, which “Polls could underestimate Mr. Trump once again”.

The method changes introduced by pollsters this year to improve the precision of opinion polls have probably made it possible to limit the extent of this underestimation, as it remains more limited in 2024 (2.75 points on average) than in 2020 (4 .1 points) and in 2016 (3.6 points).