Attempting to reconcile facts and narratives.

It has taken a while to begin to get some of the actual breakdown of voting numbers from the election to test the narratives that are floating around. Broadly, the story has been that Trump is anathema to women and minorities and that his support is primarily white non-college educated males. Mathematically that has always been incorrect in the sense that that demographic is at most 25% of the voting population. Clearly there is more going on than the narrative allows.

Correspondingly, the story has been that Clinton would triumph due to minorities, women, the poor and the wealthy. Again, it is hard to make those numbers quite work.

So what was the reality? Of course there are elements of truth to each story. From Pew:

Trump’s margin among whites without a college degree is the largest among any candidate in exit polls since 1980. Two-thirds (67%) of non-college whites backed Trump, compared with just 28% who supported Clinton, resulting in a 39-point advantage for Trump among this group.

Similarly for Clinton, part of the story is true:

Women supported Clinton over Trump by 54% to 42%. This is about the same as the Democratic advantage among women in 2012 (55% Obama vs. 44% Romney) and 2008 (56% Obama vs. 43% McCain).

So the macro stereotypes are supportable. But the interesting thing is what is going on in the details.

The media wishes to present Trump as the candidate of the uneducated. There is some evidence to support that.

College graduates backed Clinton by a 9-point margin (52%-43%), while those without a college degree backed Trump 52%-44%. This is by far the widest gap in support among college graduates and non-college graduates in exit polls dating back to 1980. For example, in 2012, there was hardly any difference between the two groups: College graduates backed Obama over Romney by 50%-48%, and those without a college degree also supported Obama 51%-47%.

But that is not the full story. It is not quite true. Indeed, not fully true at all. Sotto voce it is revealed that

Trump won whites with a college degree 49% to 45%.

So among college educated whites, Trump beat Clinton by 4%. Bet you didn't know that from the MSM. You don't hear much of Clinton being the candidate of uneducated whites. But apparently that's what the numbers are saying.

What about the gender trope? Again, there is a basis for the stereotypical claim.

Women supported Clinton over Trump by 54% to 42%. This is about the same as the Democratic advantage among women in 2012 (55% Obama vs. 44% Romney) and 2008 (56% Obama vs. 43% McCain).

But again, nuance raises its ugly head. From CNN Exit Polls, white women went for Trump by a margin of 53% to Clinton's 43%. Again, did you know from the MSM that Clinton only got 43% of the white female vote?

Finally, it has been a mainstay that Trump is a racist enemy of people of color. But did you know that Trump increased Republican share among people of color? Again, not widely reported. Trump increased Republican share of the Hispanic vote from 27% to 29% (a 7% increase) and African-American vote from 7% to 8% (a nearly 14% increase). Of course, the base remains depressingly low but those numbers once more contradict the MSM narrative.

Since white voters remain the single largest voting demographic (70%), what happens there is significant. How can there be such a disconnect between the white voting patterns and the MSM narrative? Of course, part of it is simply different voting patterns among people of color. But it is hard to make those numbers work. How can Clinton take only 43% of white women and yet get 56% of women overall?

I don't really know.

They only thing I can hypothesize is an untested observation. My impression is that African American women invest heavily in education. They are disproportionately represented in occupations that financially reward advanced education such as Government and Education (teachers). My guess is that African American women are disproportionately represented in the people of color vote and that they also have a disproportionate degree of college education for the entire people of color population. I suspect that is part of the discrepancy, but only a part.

So we are left with the dichotomy. The press maintains that Trump is supported primarily by uneducated white males whereas the numbers say he took the white educated vote and he took the white female vote.