Trump's obvious and continuing glee at having won — glee that's totally justified, of course — is clearly tempered by the asterisks that apply. The biggest of those asterisks is that he trails in the popular vote, which he falsely claimed was a function of voter fraud. But he also went out of his way to suggest that he did far better among groups that were expected to oppose him than anyone predicted.
A constant refrain of his pre-election rallies was that everyone loved him. He used to say he'd win the Hispanic vote, for example, and always pooh-poohed the idea that women would vote against him. On Thursday, he summarized the results, in a sense.
The African American community was so great to me in this election. They were so great to me. Amazing. I couldn't believe it. I started off at a low number and every week boom, boom, boom. And I got it up to a number that's higher than all of the Republican candidates for years, and it was great.The Hispanic community, I did great with the Hispanic community, higher than people that were supposed to have done well. I felt it.And is this really a big surprise? We did great with women. Can you believe it? Great with women. A couple of polls came in . . . [booing] A couple of polls came in the early states, and they said, we don't believe it. He's doing well with women. But every time I went out, I saw those beautiful pink signs, right? Women for Trump, and I knew we were gonna do well. So, we did great with women, we did great with everybody.
Let's consider those claims.
Support among black voters. Over the course of the campaign, Trump regularly claimed that his support from black voters was ticking upward. There's no public evidence that it was. Often, the sample size of African Americans included in polling was too small for the results to be statistically significant, meaning that this claim about his poll numbers ticking upward is hard to verify. There was that Los Angeles Times/USC poll that sporadically showed Trump's support among black voters spiking, but it turns out that those increases were often a function of one guy in Illinois.
In Washington Post-ABC polling, black support for Trump was 3 percent in June and 4 percent in October. In our last tracking poll, the number was 6 percent, as undecided voters were making up their minds. There was no apparent uptick week over week.
That said, our last poll showed less support for Trump than did the final exit polls. In those polls, Trump claimed 9 percent of the support from black voters, a tick above both Mitt Romney in 2012 and John McCain in 2008.
McCain and Romney, though, were running against America's first black president. In 2004, George W. Bush wasn't, and he got 11 percent of the vote according to exit polls. Of the 11 contests between 1972 and 2016, in only three did Republicans do worse with black voters. (The third was Bush in 2000 — when Bush got 8 percent.)
Support among Hispanics. This is the point at which we acknowledge that data on nonwhite voters in exit polling is a bit iffy.
This is the pattern among Hispanic voters, akin to the graph above.
Trump did about as well as Romney, and worse than McCain and Bush. He did better only than the two Republicans who faced Bill Clinton in the general.
But other polling suggests that the numbers here are wrong. One Latino pollster took issue with these results, outlining a number of reasons they were skeptical of the final numbers. (Among them, past analysis demonstrating higher margins of error among nonwhite populations.)
Those 2004 results, for example, spurred a great deal of analysis given how well they showed Bush faring against John F. Kerry. State exit polls suggest that Bush's support among Hispanics was closer to 40 percent — still high, but 4 points different.
Trump's claim about the Hispanic vote is vague, though: “higher than people that were supposed to have done well.” So we'll leave it at that.
Support among women. In this case, Trump's sort of correct. Pre-election polls showed him losing in large part because women, including Republican women, rejected his candidacy. That didn't come to fruition.
Trump still did worse with women, mind you. He had the worst performance for a Republican with women since 1996 and the third-worst performance since exit polling began. Among white women, a group that normally votes Republican, he did a bit worse than Romney, but not by much.
This was the group that made the difference. Late in the campaign, white women — largely white women without college degrees — got onboard the Trump train. Whether Trump seeing pink signs in his audiences offered a clue as to what was coming is debatable. But he did better than expected.
What's not true is that last part: “We did great with women, we did great with everybody.” He lost among women. He lost with black voters and Hispanics. He won with men, mostly white men. But Trump's goal isn't to convey accurate information, it's to cast himself in the best possible light.
Another way in which Thursday night's rally was just like another day on the campaign.