Hillary clinton was given for sure winner of the 2016 election. How it went then, we all know. And the pollsters know this above all, who since that night have lost sleep, because their polls, although accurate in detail, were essentially wrong and gross.
In the last four years, however, the survey institutes have worked hard to try not to repeat the same mistakes of 2016, taking care to better weight the samples, to introduce more non-graduates, to add the psychological variant of Trump-shy they vote for Trump but are ashamed to declare it, observing more carefully what happens on social media, making more cautious forecasts.
In particular, there are various factors that make Biden’s lead seem more solid and, so to speak, certain compared to Clinton’s. The first is its breadth: to date Biden is given between 9 and 12 percentage points ahead, while two weeks after the vote in 2016, Clinton’s advantage was between 2 and 4 points. Not only. But in 2016 there was the very unlikely circumstance that Trump won all the states in the balance, something already unlikely in his own right, something almost impossible in a scenario in which all the polls (made with the criteria of 2016 and not of 2020) they gave Trump back.
Of course, even today, as four years ago, the Democratic candidate is given ahead both nationally and in individual states. But the way in which the detections were made has changed, is more accurate and broader, and at the same time there is greater caution in making triumphalistic predictions about the victory of one or the other (remember New York Times giving Clinton the 98% chance of being elected?), Because the elections are not won by whoever gets the most votes but whoever gets the largest voters, who are only 538. The balance, with such a small number, could be quickly reversed. Trump would need two or three key states of weight (like Florida or Pennsylvania) to reopen the games and reduce your disadvantage.
It can happen, we know. But in 2016, however, we did not know