President-Elect Donald J. Trump. The only poll that mattered--the actual election--broke the way that none of the polls predicted, and America has its own seismic version of Brexit to explain.
So what happened? Why was it the case that Hillary Clinton lead in the Real Clear Politics Poll of Polls for all but a few days in late July? Only two weeks ago, Hillary Clinton also lead in almost every national poll by double digits.
Here are the most obvious explanations:
1. Turnout. In a column I wrote two weeks ago, I called this the "One Factor That Could Betray the Polls." Every single pollster has an underlying turnout model in their polls. These turnout models help them determine who the likely voters are. They are often based on past elections. And, this year, it looks like these turnout models did not pick up a surge in Republican voters in the suburbs and rural areas.
2. Focus on National versus State polls. National polls can be of limited value when determining the state by state results which determine the Electoral College. Despite the 3 to 5 National lead, many of the state polls had closed within the margin of error. Pollsters potentially overstated the strength of the results in the states, often focusing on "who was in the lead" vs. "what the statistical results" could really discern--a statistical tie.
3. Correlation. Although we often focus on state boundaries, it is easy to forget that voters in geographic areas can be very similar and they may well vote the same way. So, when Ohio broke early for Trump, it was a sign that there could be a wave of Midwestern states made up of suburban, white voters started to tilt Republican. This would explain Trump success in Wisconsin, Michigan, and Pennsylvania.
The polling profession has been in a bit of crisis after Brexit, and this result will only further cast doubt on the current way polling is done. These three factors are going to be part of the story as we look back on the Election of 2016.