2016 Postmortem
Related: About this forum538: Clinton Gains, And The Polls Magically Converge
irst things first: Hillary Clinton has a 70 percent chance of winning the election, according to both the FiveThirtyEight polls-only and polls-plus models. Thats up from a 65 percent chance on Sunday night, so Clinton has had a good run in the polls in the final days of the campaign. Clintons projected margin of victory in the popular vote has increased to 3.5 percent from 2.9 percent.
Well continue to collect polls through early Tuesday morning, at which point well update the model for the last time and publish a more philosophical overview on the race. But Im not sure how much more data were really expecting most of it will probably just be state and national tracking polls that run one last update.
As a lot of you noticed, Nevada, North Carolina and Florida flipped from red to blue over the course of Monday. We dont think thats a particularly meaningful metric, because the forecasts are probabilistic Clintons chances of winning Florida increased to 54 percent from 48 percent, for instance, which is nontrivial but not an especially large change. Still, we know its something a lot of readers follow. Its unlikely that any further states will flip to Clinton in our final forecast, as shes too far behind in Ohio, the next-closest state.1 Its possible that Florida and North Carolina could flip back to Trump by tomorrow morning, though probably not Nevada, where Clintons lead is a bit larger....
Seven of the 19 polls have Clinton leading by 4 points; another four have her ahead by 3 points, then we have a smattering of 1s, 2s and 6s along with two pollsters, IBD/TIPP and Los Angeles/USC Dornsife, who still have Trump ahead. (We admire the L.A. Times poll for not changing its methodology in midstream, even though the poll has its issues.) On average, Clinton leads by 2.9 points in the polls, although the highest-rated pollsters2 have her a bit higher at 3.8 points, on average. As is usually the case, the range of national polls closely matches the FiveThirtyEight popular-vote forecast.
Its worth raising an eyebrow, though, when the polls (other than the L.A. Times) show a range this tight at the end of an election, especially given that theyd diverged so much earlier in the campaign. That probably reflects some degree of herding for instance, because pollsters stick surveys that seem to be outliers in a file drawer rather than publishing them. So the tight range of polls shouldnt be taken to mean that everyones figured exactly how to poll this challenging election just in the nick of time. Still, the polls clearly agree that Clinton is the favorite, and perhaps has a slight wind at her back for Election Day.
http://fivethirtyeight.com/features/election-update-clinton-gains-and-the-polls-magically-converge/