As I've observed in a number of posts, the relationship between education and voting has been gradually shifting--college graduates used to be more likely to vote Republican, compared to non-graduates with similar incomes, but now are more likely to vote Democratic. Although the shift hasn't been completely steady, it's been gradually progressing over a long period of time. Amory Gethin, Clara Martinez-Toledano, and Thomas Piketty have a paper looking at the relationship in 21 democracies, and find that that the same change has occurred in almost all of them. They propose that it's actually a bit more complicated than that, as in multiparty systems there are now different kinds of left and right parties--those that emphasize social issues and whose voters are more divided by education and those that put more emphasis on economic issues whose voters are more divided by income. In the United States, we have a two party system, but there have been a few third-party candidates for president who have received a significant number of votes. I calculated the average education (measured by highest degree, ranging from 0 for not a high school graduate, to 4 for a graduate degree) for voters for different candidates in elections from 1968 to 2016 (limiting it to whites, since non-whites have voted heavily Democratic regardless of education).
In the earlier elections, Republican voters were generally more educated than Democratic voters, but since 2000 Democratic voters have consistently been more educated than Republicans, and the gap in 2016 was the largest over the whole period. But my main interest is in the third-party candidates. Supporters of George Wallace were less educated than supporters of either major-party candidate, while supporters of John Anderson and Ralph Nader were more educated. Supporters of Ross Perot were just slightly less educated. Anderson's supporters stand out for their relatively high education--the gap between his supporters and Reagan's (or Carter's) was bigger than the gap between Clinton's and Trump's, and despite the large increase in overall educational levels, it wasn't until 2012 that any major-party candidate's supporters had as much education as Anderson's.
The GSS makes an effort to repeat questions, but it didn't have as many in the early years. In order to get a sense of why more or less educated people might have been attracted to these candidates, I looked at a series of questions about how much confidence you have in "the people running" certain institutions. I chose those mostly because they have been included consistently, but partly because I am interested in the general issue of confidence in institutions. I considered the following institutions: "banks and financial institutions," "major companies," "organized religion," "education," "organized labor," "the press," "medicine," "TV," "the US Supreme Court," "scientific community," and "the military." That makes a lot of numbers, so I'll just summarize them by giving the ones on which the people who voted third party candidate were more or less confident than the people who voted for either major party candidate. Small differences are in parentheses.
Wallace: less confidence in organized religion, the press, education, supreme court, (science);
more confidence in the military.
Anderson: less confidence in (business), religion, (education), TV, military.
more confidence in supreme court, science.
Perot: less confidence in religion, education, supreme court.
Nader: less confidence in business, religion, (medicine), supreme court, military.
More confidence in science.
Supporters of all leaned towards the less confident side, which is reasonable: people who are generally discontented with the way things are would be more likely to support third-party candidates. Another thing they had in common was less confidence in organized religion--that surprised me. Anderson's supporters had more confidence in science, which anticipated a divide between Democrats and Republicans that opened up later (see this paper by Gordon Gauchat). The greater confidence in the Supreme Court may reflect politics (at the time, the court was generally seen as liberal) or a more general tendency to value expertise.
This analysis wasn't as illuminating as I had hoped, but I noticed some potentially interesting changes between supporters of the Democrats and Republicans. I may write about those in a future post.