Even as Donald Trump begins to form his administration, people are still wondering why the nation's top polling organizations failed to spot his surge in the campaign's final weeks. The reasons they were wrong are important to any retrospective analysis of the election and offer a major lesson for future polling.
But more important still is that we recognize that our failure to follow the opinions of the white, high school graduate men who delivered the election to Trump is a national blind spot. It is akin to the old days when we did not recognize gender gap as political reality.
Donald Trump beat Hillary Clinton by winning white, high school graduate men in the last three weeks of the election. It was this late surge - more than any other factor - that caused his victory and enabled him to break up Hillary Clinton's "blue wall" of the Democratic states of Ohio, Pennsylvania, Michigan and Wisconsin.
According to the Fox News poll, Trump led among white high school graduate men by 14 points three weeks before Election Day. Two weeks out, he doubled his margin to 30 points. With only a week to go, his lead in this demographic swelled to 40 points. And according to exit polls on Election Day itself, he beat Clinton 67-20 - by 47 points - among this demographic. Similarly situated white women also went for Trump but by smaller margins.
How did the polling firms miss it?
They focused obsessively on the black vote, the Latino vote, and the female vote, but not at all on the white, high school educated vote. It just didn't occur to them. It was not in their lexicon of political trends. When these voters began to tip toward Trump heavily, they never noticed it.
The pollsters did not include this demographic in their data and did not report it to the journalists who recounted the survey results. Methodologically, they probably lumped white high school men in with all other high school grads, whether black or Hispanic or white or male or female. So they could not spot the trend among this crucial group.
Because these voters were not black or Latino or female or (necessarily) young, they didn't count. They were 20 percent of the American electorate, but they were never entitled to their own subset in the polling. Their votes were just churned into the national trends.
And because the pollsters and media didn't follow the sharp tilt of white high school men toward Trump in the final three weeks, they did not see any reason to commit resources to tracking in Michigan, Wisconsin or Pennsylvania. The election would be decided in Florida, Ohio and North Carolina. The Rust Belt states wouldn't matter, because they were Clinton's "blue wall."
So now white men with high school educations have finally earned their place in the sun. We will, hopefully, track their views in polls and focus groups and will understand that their power is at the core of the emerging identity politics of America.