A House Divided: US Election Demographics

As the global shock of Donald Trump's election transitions to coping with it, data are emerging that remind us how markedly divided the country is geographically, demographically, and politically. The NYTimes put together exit poll data from nearly 30,000 people, including phone interviews with mail-in and early voters. These data obviously capture voters, so there is no issue with figuring out who is a likely voter. There may still be some non-response bias, but the in-person interview technique outside polling places almost certainly reduces that considerably. Anyway, here are three pieces of information that seem noteworthy to me: (1) a majority of men voted for Trump, while a minority of women did, but it was a pretty high minority of women; (2) the less educated you were, the more likely you were to vote for Trump, but Trump still grabbed almost half of white college graduates; and (3) as expected from the electoral maps and discussions ahead of time, the farther away you were from a city center, the more likely you were to vote for Trump.

The data also suggest that the win for Trump was as much about people not turning out to vote for Clinton, as it was about people voting for Trump. He got about the same number of popular votes as did Romney, but she got fewer votes than did Obama. Her team was worried that this might happen, and it did. Nonetheless, she still received more popular votes than did Trump, even though fewer electoral votes. This has happened five times in U.S. history--three times in the 19th century and twice in this century (Bush and now Trump).