Some pollsters also wonder whether the problems may recede when Mr. Trump is not on the ballot. But he seems unlikely to be the only cause of errant polling, given how badly many congressional surveys missed the mark this year. In Arizona, Georgia, Maine, Michigan, North Carolina and South Carolina, the final publicly released Senate polls fell short by more on average than the final presidential polls.

A separate set of changes may involve how the media present polling and whether publications spend as much money on it in the future. “The media that sponsor polls should demand better results because their reputations are on the line,” James A. Baker III, the former secretary of state, wrote in The Wall Street Journal this week.

Among other questions, editors are grappling with the best way to convey the inherent uncertainty in polls.

Mr. Silver, then an independent blogger, created a breakthrough in 2008 when he began writing about every available poll, focusing on the swing states and talking about the probability of one candidate beating another. Before that, most publications had focused on national polls and largely ignored those done by competing publications.

The New York Times published Mr. Silver’s blog, FiveThirtyEight, from 2010 through 2013, and it is now part of ABC News. Other publications have since taken a similar approach, creating their own probabilistic models.

Mr. Silver and others have tried to emphasize the uncertainty in polls, by giving both candidates in a race a percentage that reflects their likelihood of winning. But many people seem to struggle to make sense of these probabilities in a one-time event like an election. They see that a candidate has a 71 percent chance of winning, as FiveThirtyEight gave Mrs. Clinton in 2016, and incorrectly think it is akin to a guarantee.

This year, FiveThirtyEight used a dot to portray each percentage-point possibility, stressing that either candidate could win.

Source: New York Times

Powered by NewsAPI.org