Was Donald Trump's large victory anticipated by the polls?

It was expected to be one of the closest elections in history. So much so that specialists in American political life had warned that we would potentially have to wait several days to know the name of the new elected president of the United States. But these announcements were not verified and Donald Trump was declared the winner at the end of an election night without the slightest suspense, Wednesday November 6.

After underestimating the votes in favor of the Republican candidate in 2016 and 2020, were the pollsters still wrong in 2024? The superposition of the averages of the latest polls shows that in reality, the Results are consistent with the margins of error of voting intentions. Franceinfo explains why with three graphs.

Both candidates were well able to win the popular vote, according to polls

Surveys are often presented in the form of simple scores. But this way of presenting their results is very partial. Surveys do not produce precise results, but rather ranges of values ​​that can be trusted based on a certain degree of confidence reflected in the margins of error. To compare the results of the election with the last polls published before the vote, franceinfo calculated the margins of error of the opinion surveys published during the last seven days before the election, on November 5.

It is very clear that the ranges of possible scores for each candidate overlap. Kamala Harris was credited with a score ranging from 46.4% to 51.6% of voting intentions for the popular vote at the national level. And Donald Trump with a score of 45.7% to 50.9%. According to these averages, the candidates were therefore both able to win.

The popular vote for Donald Trump is within the margin of error in the latest polls

During the last two presidential elections, pollsters had significantly underestimated the votes in favor of Donald Trump, as reported the political analysis site 538. But this year, the pollsters seem to have gotten it pretty right. This is at least what is shown by the superposition of the results of the ballot (still provisional on Wednesday evening) with the margins of error of the polls from the weekend of November 2 and 3.

As the counts stand, which were not final when we produced these graphs, Donald Trump is credited with 51% of the votes at the national level (compared to 47.5% for Kamala Harris), according to the AP press agency, a score within margin of error of six of the last eight polls published before the election.

The final results can still come close to the ranges given by the last two pollsters (Marist and Research Co.). “There remain a large number of ballots to be counted in California, Washington State and Oregon, three Democratic states. Donald Trump's lead should therefore narrow”explains Mathieu Gallard, research director at the Ipsos polling institute, interviewed by franceinfo.

Pollsters did not rule out a large victory in key states

But the American presidential election is not decided by direct universal suffrage, based on the popular vote. To win the race for the White House, you must obtain a majority of the electors (at least 270) who are distributed between each state according to the size of their population. In this context, the victory in several of the seven pivotal states, the famous swing states, was decisive. They could switch to one camp or the other.

In the week before the vote, the ranges of scores given to the two candidates overlapped in all the key states. According to the latest polls, it was therefore possible that the two candidates shared these swing statesbut it was also possible that a single candidate would win the majority of them, as recalled by the New York Times on the eve of the election.

It is this last scenario which occurred on the night of November 5 to 6. By winning North Carolina, Georgia, Wisconsin, Pennsylvania and even Michigan (Wednesday afternoon after a long count), Donald Trump ensured a clear and quick victory.