After seeing this post up on how high information levels and education may lead to political polarization, I wanted to revisit the GSS data on political moderation and independence in light of educational attainment and intelligence. For the later I used the proxy of a score on a vocabulary test which has a 0.70 correlation with general intelligence. My question was this: what are the different effects of intelligence and education on ideology and partisanship?
To answer this question I looked at two response variables, POLVIEWS and PARTYID, which measure ideology from very liberal to very conservative and partisanship from strong Democrat to strong Republican. I amalgamated “leaners” so that in the middle I had moderates and independents left. For the vocab test I used WORDSUM. The scores have a value from 0 to 10, out of 10. I combined the 0-4 interval because the sample size there was small. Finally, I limited the sample to non-Hispanic whites after the year 2000 to eliminate some background confounds (e.g., minorities tend to be way more Democratic, all things equal).
I generated some area graphs. First, I looked at proportions of each ideology or party in a particular category. For example, the percentage liberal, moderate, and conservative who get a vocab score of 5, or have high school education. Then, I controlled for education and looked at vocab score. Specifically, I limited the sample to those who had only a high school diploma, and then those who had a university degree and higher. These two classes had large sample sizes. Then I looked at how ideology and party varied by vocab score.