Donald Trump’s clear victory in the presidential elections has not only put the Democrats on the grill, but pollsters and forecasting gurus. The first two names targeted are those of the historian Allan Lichtman, considered the “Nostradamus” of predictions for having guessed all the results of the presidential elections in the last forty years, except that of 2000, decided at the last vote, between Al Gore and George W. Bush, and the forecasting ‘guru’, the political scientist Nate Silver. Both had predicted Kamala Harris’ victory. Lichtman publicly admitted that he was wrong, but calling his rival into question: “Unlike Nate Silver – declared the professor – who will try to avoid having to explain why he didn’t understand what was happening, I admit that I was wrong”. Silver, at the helm of America’s most followed polling site, FiveThirtyEight, had said for weeks that the challenge between Harris and Trump would be like “flipping a coin”, given the great uncertainty, but on the eve of the vote he had indicated the vice president as the probable winner.
In reality, not only did Harris not win, but there wasn’t even a head-to-head, as Trump won all seven key states and also took more votes than his opponent in the global count. Even the pollsters have collapsed: for months the various institutes have stunned millions of Americans with data on state-by-state challenges and on the popular vote, portraying a country that was actually going the other way.
From Ipsos/Reuters to Research Co, from Emerson College/The Hill to the Washington Post, from the Des Moines Register/Mediacom to the New York Times/Siena College, no one has captured the true state of the country. In an MSNBC report a few days after the polls opened, around twenty “experts” had made their prediction: journalists and podcast hosts had no doubts, “Trump would declare victory in advance to muddy the waters, but then Harris would have won.” According to some, “with a blue wave of votes”.
Even the analyzes published in the media appeared a bit messy, like the one, published on October 8, by the analyst Nate Cohn who in the New York Times had done a long analysis on the vote, but got the calculations wrong, starting from the differences between two percentages of voters: for him between 55 and 41 there were 13 points. A more attentive person would have indicated 14 as the correct difference. But the confusion and sloppiness of the analysis seemed like the tip of the iceberg of the polling business: in every election, millions of Americans looked at the numbers and relied on them. Each new survey attracted clicks, but ended up distancing the “readers” from the real perception of the “voters”.
What also made the fiasco of the polls more sensational was the fact that this is not the first time that the institutes have made predictions wrong: in 2016 the New York Times announced, on the eve of the vote, that Hillary Clinton had a 91 percent chance of win against Trump’s 9 percent, who then won with a landslide like this year. The tycoon’s electorate has always been underestimated.
Even four years ago, Joe Biden was given a lead of 9 points over Trump by the FiveThirtyEight site, which tracked the average of the polls, and then won with about half the expected gap. Furthermore, the Democratic candidate was clearly ahead in Arizona and Georgia, where he won by a few thousand votes, and ahead in North Carolina and Florida, where he then lost. This time, precisely because of the size of the result and the resounding flop, Americans may begin to no longer take the polls as scientific documents. Analysts like Silver will be viewed with less confidence. And perhaps the Nostradamus of elections, Professor Licthman, will not be asked to point to a winner based on his prediction model. Or if it happens, Americans are warned: don’t trust too much, because the so-called American experts probably know as much as you do.