First, Heartiste notes a commenter at Steve's pointing out what I should've emphasized:
Which of these things is not like the others?A couple of clarifying comments. Steve's subsequent post looked at work from professor George Hawley who asserted that "the relationship between marriage and voting declined" in 2016 from 2012.
Single white women: 39.0
Married white women: 61.4
Single white men: 59.0
Married white men: 61.0
Maybe--Trump was more Fishtown and less Belmont than Romney--but the correlation he looks at is at the state level, not at the level of individual voter. I've not run the numbers, but I suspect that at the state level the relationship between IQ and voting Republican increased between 2012 and 2016. The upper Midwest states, where Trump outperformed Romney the most, do pretty well on measures of intelligence.
Given Trump's much stronger performance with working-class whites, modestly stronger performance with non-whites, and poorer performance with professional whites relative to Romney's, however, it's almost certain that the average IQ of Republican voters decreased between 2012 and 2016 even though the average IQ of the states that went for Trump was higher than the average IQ of states that went for Romney (clever sillies aren't going to save the West, Alt Knights will).
The officially commissioned exit poll released on election night shows a marriage gap that is narrower than what the R-I results show. I wish I'd saved all the initial releases on November 9th. I did happen to screen shot a few in the course of doing a little post-election analysis before they were changed again and again and again, presumably to adjust to official electoral results as those trickled in for several weeks after the election. Of the four exit poll tables captured in this post from November 12th, the results for every single one subsequently changed.
In other words, there's reason to be skeptical of the commissioned exit poll. In any event, R-I's results provide another similar but not identical data set.
Heartiste, pointing out Trump's strong performance among married Jewish men, cautions that this might more accurately be described as a strong performance among married Orthodox Jewish men on the presumption that a lot of secular, irreligious ethnic Jews identify as having "no religion" rather than "Jewish".
It's hard to know how things shake out precisely from survey to survey, but the GSS--when asking "What is your religious preference? Is it Protestant, Catholic, Jewish, some other religion, or no religion?"--shows that only 36.1% of those who identify as Jewish say they "know God exists". The corresponding figures for Protestants and Catholics are 74.9% and 62.2%, respectively.
In other words, barely one-third of self-described Jews--self-described in response to a question about religion rather than ethnicity or cultural heritage, no less--are firm theists. Some number of secular ethnic Jews do not identify with any religious tradition, but a not insubstantial number of them--the majority if I had to guess--identify as Jewish.
GSS variables used: RELIG(1,2,3), YEAR(2000-2014), GOD(6)