Murc’s law, racial unmarked category edition

A revealing sub-genre of American journalism is the “why are so many [Black, Hispanic, Jewish, Asian etc.] people voting for Trump?” piece.

Here’s the lede from a typical one by Matt Yglesias:

President Donald Trump’s current general election polling is dismal, putting him down about 9 percentage points in general election polling averages, which is far too large a deficit for individual state peculiarities to matter. That slide in the polls includes the evaporation of modest gains with African American voters that were visible last year, and substantial defections from the large bloc of older white voters who were very solidly in Trump’s camp in 2016.

But the decline has not been seen across the board. As Domenico Montanaro reported in his writeup of NPR’s polling on the race, “the one group Biden continues to underperform with slightly is Latinos — 59% of Latinos said they’d vote for Biden over Trump, but Clinton won 66% of their votes in 2016.”

Trump’s relative resilience with Latino voters can be easy to overlook because he is losing these voters by a large margin (39 points, according to the New York Times). Still, he is losing them by less than he did in 2016, which is strange at a time when his numbers are otherwise falling.

(Related side issue that I’m going to write about separately: it’s interesting that as of 2012 Yglesias considered himself “just another white dude,” rather than a member of an ethnic minority group in any meaningful way. He has a Cuban grandfather while his other three grandparents are Eastern European Jews. In the 2020 version of America, that seems like a fairly liminal or contingent version of “white.” Anyway I wouldn’t be surprised if he doesn’t feel somewhat differently about the matter now.)

Now if we’re talking about white people, it doesn’t seem too probable that Trump could be losing some white demographic to Biden by 39 points, yet still have that level of support be described as surprisingly robust. Oh wait, we can actually test this hypothesis. From the very same article Yglesias quotes that found Biden leads Trump by 39 points among Hispanics:

Most stark may be Mr. Biden’s towering advantage among white women with college degrees, who support him over Mr. Trump by 39 percentage points. 

Now to be scrupfair, Yglesias is arguing that, while Trump’s support has collapsed among white women with college degrees, it has remained steady among Hispanics. But the point is that the steady degree of support for Trump among Hispanics is exactly the same as what is described (accurately) by the Times as Trump’s catastrophically bad standing with college-educated white women!

Here’s another example from Vox (which I generally think of as one of the best political publications in the country — that it too is so prone to this journalistic tic is telling):

As a general rule, black Americans do not support Donald Trump. According to a recent Wall Street Journal/NBC poll, Trump enjoys just a 14 percentapproval with black Americans, while roughly eight in 10 black voters say they’re “uncomfortable” with his 2020 run for reelection. In a poll of roughly 800 black registered voters conducted by BlackPAC, Trump had a -59 percent net job approval rating.

But within those numbers is another story — a stark gender divide.

Roughly 24 percent of black men polled by WSJ/NBC approve of Trump’s efforts while in office (72 percent of black men disapprove), but that number plummets to 6 percent when black women are asked the same question.

Exit polling from the 2016 election shows that while 13 percent of black men voted for Donald Trump, just 4 percent of black women did (in Pennsylvania, that number dropped to 1 percent). For comparison’s sake, a majority of white men and white women voted for Trump, as did 32 percent of Latino men and 25 percent of Latino women. And this “black gender gap” isn’t new. Exit polls for 2008 and 2012 show that more black women voted for Barack Obama than did black men. In fact, the last time more than 10 percent of black women voted for a Republican presidential candidate was 1996, when 14 percent of black women voted for Bob Dole (to compare, 22 percent of black men did the same).

Again, the real lede here is that non-whites oppose Trump by margins that would be considered completely overwhelming if we were talking about the political views of white people. If 72 percent of a white demographic disapproves of the president, journalists have no trouble concluding that those white people hate themselves some Donald Trump. And that’s because 72 percent disapproval represents a massive repudiation in normal democratic terms. But normal democratic terms don’t apply to non-white voters and Trump, because apparently those are the only voters who have real moral agency when it comes to our home-grown reality TV fascist.

White people amirite?

Maybe these sorts of stories should actually be about how the only reason Trump got elected, and remains even slightly viable as a politician, is because so many white people vote for him. Because the rest of America has always rejected him by overwhelming margins, and continues to do so.