Nate Silver, in a commentary on FiveThirtyEight, undertook a statistical analysis of who voted for whom in the recent US election. To do this, he compiled data from all 981 US counties that contained 50,000 people or more and sorted those by the proportion of the population that had completed at least a four year university degree. The surprising result was that in 48 of the 50 most well educated counties (averaging 51.4% with bachelor’s degrees), Hillary Clinton actually improved on Obama’s 2012 performance, given as the margin of victory, increasing it by 8.5% on average. Conversely, in the 50 least well educated counties (averaging 13.3% with bachelor’s degrees), Clinton lost ground relative to Obama in 47 of the 50 counties, falling behind by an average of 11.3% on average. The latter were where Donald Trump won the presidency.
One question which Silver asks is: Could it be income levels rather than education levels which drove the changes in support? It is difficult to determine this because there is a very strong correlation between education and income. This correlation is very strong, with a correlation coefficient of 0.69 for share of the population over 25 with a bachelor’s degree. Despite this, Silver has found some areas that have high levels of education, but average or below average levels of income. If education is the driver for differences in support, then you would expect to see Clinton hold Obama’s lead or improve upon it. It turns out that in most places where education level was high but wages average or lower, Clinton improved on Obama’s performance.
There are also counties in which residents have high incomes, but education levels are not particularly high. In counties like this, support for Clinton tended to drop dramatically when compared with the support given to Obama in 2012.
The interpretation of these data are difficult, because:
- Education levels may be an indicator of a liberal world view
- Educational attainment may be a better indicator of economic well-being than household income.
- Educational attainment probably is inversely proportional to suspicion of outgroups
- Education levels are strong indicators of media consumption habits
- Trump’s approach to the campaign – relying on emotional appeal rather than policy – may have appealed to those with lower education levels.
It is no wonder that Trump stated that he loves the poorly educated, perhaps because he knew he could appeal to them with any sort of emotional bullshit.
I also suspect that is why conservatives are not in the least bit interested in improving the education of the general populace, so they keep schools as poor as possible, while making it as difficult as possible to go to university. It garners them more votes because uneducated people are more easily gulled.