Good read - Dangerous idiots: how the liberal media elite failed working-class Americans

[whatthehey]

1. Somebody's screwed up the numbers then

Because if we believe this, which is just the first of many polls a quick search found on the "education gap":

http://www.bloomberg.com/politics/articles/2016-08-12/education-level-sharply-divides-clinton-trump-race

Then that cannot possibly match a 44% graduate rate among his supporters. Let's do the math.

Clinton leads Trump 59/34 among graduates. Let's scale those up to get rid of undecideds/3rd parties. We need to do this because Gallup isn't looking at levels of support but education level among the total group of Trump supporters. So we can imagine that, gun to head and presented with a binary choice, 36.6% of graduates would go Trump.

Nongraduates? Trump leads 52/42. Do the same scale up to total population and we get 55.3% of them.

Now we need to translate that into how many Trump supporters have degrees, rather than the other way round. Imagine, to make translation into percent easy, there are only 100 people. 29 of them have degrees, 71 don't. Trump gets 36.6% of the 29 and 55.3% of the 71. Or, rounded, 11 and 39. Of his total 50, only 22% then would have degrees.

It's not much better if we use the percentage of whites, almost all his support, who have degrees. Then he'd have 36.6% of 33 and 55.3% of 67 for 12 and 37 or 24.4% degreed.

It's easy to quibble that extrapolations may be a bit off because there really is no gun to head binary choice, or that racial and income data skew the calculations a bit, but think of the magnitude. We don't have to shift the numbers a bit, we have to pretty much double them to get Gallup's hypothetical degreed Trumpites. Every poll I've seen that covers education level shows the same graduate gap. There is simply no way to reconcile that aggregate look with a Trump support that is 44% graduate.