Using CNN's exit polls, I decided to do a little bit of work using Excel/R to investigate the relationship between how college whites vs. non-college whites voted as well as how white men vs. white women voted in the states in which exit polls existed for 2016.
If I have more time, I will try to do something similar for previous years, but I think users may find this data interesting and it has some interesting data to sift through within it.
Note that in this plot:
x-axis = difference in GOP margin between how white men and white women voted (white men Trump margin - white women Trump margin)
y-axis = difference in GOP margin between how non-college educated whites and college-educated whites voted (college white margin - non-college-white margin)
Also if anyone else wants a CSV file to toy around with the data themselves, shoot me a PM and I can email it to you.
A few trends I notice immediately is that Trump clearly struggles far less with college whites in Ohio, Iowa, and Florida than the vast majority of states. Also, with a calculated r value of -.033, there seems to be no connection between the gender gap and the education gap among whites in different states - as seen by outliers like Iowa and New Jersey, for instance.
I'm curious what other trends people notice here - I'll probably have some more fun with this later (and make the graph nicer), but it seems like there are some interesting trends to be seen from this data as it is.
Also, if you're having problems seeing the image, just go directly to the url,
https://postimg.org/image/o8o6k9x89/