Pew Research: Republicans More Knowledgeable Than Democrats

[muon2]

Nathan, Torie, and bgwah make sound points. The issue involves a complex set of demographic variables, as the following graphs show:



Whites without high school degrees are less Republican than white high school graduates, but in each education category beyond that, voters are more Democratic. Richer people within any education category are more likely to vote Republican.

Interesting that poor post grads are even more reliably Dem than the poor w/o HS. Presumably the rural poor shift enough away from the urban poor to make that so. I'm guessing that the income axis is simply the five quintiles.

[muon2]

I'm guessing that the income axis is simply the five quintiles.

http://themonkeycage.org/blog/2012/03/23/voting-patterns-of-americas-whites-from-the-masses-to-the-elites/#more-16255

You must be logged in to read this quote.

Those are unusual categories for income. The first three correspond reasonably well to the three lowest quintiles so each would have about 20% of the population. The 75 K is a little high for the third quintile (65 K would be better), and adds about an additional 5% or more to the sample there. If that were the only issue, I wouldn't have much to quibble about.

However, If I wanted equal numbers of people in each data point, the next cut shouldn't be at 150 K. That leaves only about 5% of the population in the uppermost bin and about 30% in the next to highest bin. The cut should be at about 100 K, so it's not close to a balanced sample. The effect is that any attempt to draw conclusions from the slope of the line at the high-income side will be skewed by the shift of population from the top data point into the next-to-top point.