When the results of the Presidential election became clear late in the evening on November 8th, 2016 and political pundits started tearing up on the air, there was a rush to blame someone. In the days, weeks, and months following, this blame seemed to be placed in large part on the shoulders of the polls and pollsters who fooled American into thinking a Hillary presidency was a lock. The New York Times declared that “Data Failed Us”.
However, with a little time and distance, some of my favorite data guys at FiveThirtyEight decided to take a look at the accuracy of the polls for the first time since the 2016 primaries. The results showed polls are as accurate as they have ever been when it comes to predicting election results, more specifically, the 2016 presidential election polls were as accurate as polls of presidential elections have been on average since 1972 and polls of elections since the 2016 election have been more accurate than average. Even the state level polls that were below average were within the normal range of accuracy.
It seems whenever there is an upset to popular (read liberal) sentiment, like the election of Donald Trump or Brexit, there is an outcry that inaccurate polling is to blame. While there are issues with polling, like the statistical bias towards Democrats in 2016 or declining response rates, overall there no polling crisis like the media would have you believe.