Since Tuesday night, there has been a lot of handwringing about how the media, with all its fancy analytics, failed to foresee Donald Trump’s victory. …
Whatever went wrong with the polls in this country, they inevitably colored perceptions. …
The prediction models didn’t help things. On Tuesday morning, FiveThirtyEight’s “polls-only” prediction model put the probability of Clinton winning the presidency at 71.4 per cent. And that figure was perhaps the most conservative one. The Times’ Upshot model said Clinton had an eighty-five per cent chance of winning, the Huffington Post’s figure was ninety-eight per cent, and the Princeton Election Consortium’s estimate was ninety-nine per cent.
These numbers had a big influence on how many people, including journalists and political professionals, looked at the election. Plowing through all the new polls, or even keeping up with all the state and national poll averages, can be a time-consuming process. It’s much easier to click on the latest update from the model of your choice. CONT.
John Cassidy, The New Yorker