AnalysisThese Biden-Trump polls that don’t say much

43-44%, 37-37%, 41-43%: all these polls which predict a very close presidential election between Joe Biden and Donald Trump are making the United States dizzy, especially when they are overanalyzed in the media. But what are they really worth? Probably not much to predict the final result next November.

It has been months since Donald Trump had a certain lead in most polls, but in recent days there have been warning signs of a certain decline, of fatigue, in support for the ex-president republican. Could his criminal trial in Manhattan, which makes headlines on a daily basis, have something to do with it? Too early to say.

Still, former President Trump’s lead over President Biden has diminished in recent months, and the most recent polls show a neck-and-neck race, as Joe Biden gains ground thanks to support from key demographic groups.

According to FiveThirty Eight, a poll aggregator which compiles and weights opinion surveys, the average of polls gives 41.04% of voting intentions to Donald Trump, compared to 40.6% for Joe Biden. Suffice to say that we are well below the margin of error.

Armed with his crystal ball, Larry Sabato, who recently celebrated the 25th anniversary of the founding of his Center for Politics at the University of Virginia, warns those who take these proliferating polls at face value. They can show over time what’s happening in major subgroups, the demographics that matter in American politics, and whether a subgroup is clearly shifting its preferences, but in terms of telling us who’s going to win , just flip a coin.

A propaganda tool

These polls are mainly used for political propaganda purposes to promote progress in voting intentions or to mobilize donors for candidates who are lagging behind. : \”Look at these numbers, of course, they are set in stone.\” They can’t change, the elections are over. And then all of history shows us that that wasn’t true”,”text”:”Whichever side is favored at the moment is going to look at that poll and say, “Look at those numbers, sure, They are set in stone.\”They can’t change, the elections are over. And then all history shows us that it wasn’t true”}}”>Whoever’s favorite side is at the moment will look at this poll and say, “Look at these numbers, of course they’re set in stone.” They can’t change, the elections are over. And then all history shows us that it wasn’t truenotes Larry Sabato.

A little less than six months before the November electoral meeting, the margin of error of the polls would be, according to him, around 6%, while the week preceding the election, it was only 1.7%.

Speaking of errors, in 2016, the polling houses were clearly wrong about the real outcome of the vote, with the victory of Donald Trump over Hillary Clinton.

Even in 2020, they suffered some embarrassment, because even though Joe Biden won by more than 7 million votes over Donald Trump, says Larry Sabato, it really came down to 43,000 votes in the tightest key states . million votes cast.”,”text”:”Forty-three thousand votes out of more than 160 million votes cast.”}}”>Forty-three thousand votes out of more than 160 million votes cast.

Fictitious polling houses?

Today, many polling institutes make projections with a sample of 600 or 700 people, sometimes even 400 or 500, which does not give a good snapshot of the situation. Larry Sabato deplores that once again the media are extrapolating too much from this less reliable data. The press lets them do it because they also like rumors. I don’t mean to be cruel, but there are a lot of people in the journalism profession who don’t understand numbers at all.

The other phenomenon that we observe when looking at these probe shots is the proliferation of polling houses which seem to come out of nowhere and which, very often, have opinion surveys which are commissioned by partisan groups, although they deny this by proclaiming themselves to be non-partisan. These results are then shared in the echo chambers that are the various partisan publications, then taken up by the offices of the national parties.

For a while we used survey averages and thought that that eliminated the problem, that we could eliminate the error to some extentexplains the founder of the Center for Politics at the University of Virginia.

The parties understand and therefore know that if you create four or five fictitious pollsters that cannot easily be classified as Democratic or Republican, the average will be affected, and they do this especially in the key states of the political spectrum.

Why survey especially six states?

President Joe Biden and his Republican opponent are engaged in a tight presidential race that could indeed depend on how the votes of the electors are cast in six key states: Arizona, Georgia, Michigan, Nevada, Pennsylvania and Wisconsin.

All of these states went for Joe Biden in 2020, but the number of voters of diverse ethnicities and ages could come into play in the next election. Votes which do not yet seem to have been definitively mobilized, several months before the election.

What about the Kennedy factor?

In such a tight race, Robert F. Kennedy Jr.’s candidacy has the potential to swing the election, although it is unclear in whose favor. Earlier polls suggested that Kennedy Jr., scion of the nation’s most famous Democratic family, could attract more votes that would have gone to Joe Biden, but more recent polls suggest he would also hurt Donald Trump.

On this point, the political scientist from the University of Virginia tempers the general perception a little. He doesn’t believe Robert F. Kennedy Jr. will take as many votes from Mr. Trump as he did from Mr. Biden.

: Jill Stein, Green Party candidate, Cornel West, African-American independent candidate who might only get 1% or less of the popular vote, but where he gets it might really matter in some states, and of course RFKJr.”,”text”:”Donald Trump has three trump cards in his game for Joe Biden to lose essential support, however modest, underlines Larry Sabato. These are three candidates likely to take more votes from the Democrat: Green Party candidate Jill Stein, African-American independent candidate Cornel West who might only get 1% or less of the popular vote, but Where he gets it might be of real importance in some states, and of course RFKJr.”}}”>Donald Trump has three trump cards in his game for Joe Biden to lose essential support, however modest, underlines Larry Sabato. These are three candidates likely to take more votes from the Democrat: Jill Stein, Green Party candidate, Cornel West, African-American independent candidate who might only get 1% or less of the popular vote, but Where he gets it might actually matter in some states, and of course RFK Jr.

“ Wait and see »

Of course, everything can change, as seen in recent polls. Joe Biden has between now and the presidential election a much richer reservoir of negative publicity to use against Donald Trump, thanks to the latter’s incendiary statements and ideas, or even his legal setbacks.

But trends evolve, a bit like tectonic plates, at their own pace.

So should we look at these polls now? Not really, except for political junkies. Before October, this will not be of much use, because the Americans’ choice is far from being crystallized.

We will have to wait and see what the question at the ballot box will be: the price of gasoline, the situation at the border, the war in the Gaza Strip? Or another surprise?

The only (almost) sure thing is that of Joe Biden and Donald Trump, only one will be president…