I don’t know the methodology used to create these distributions, but this graphic from a recent Economist article on the political effects of social media shows how Democrats and Republicans have overlapped less and less in the last 25 years:
I would really like to know the method behind creating these shapes. For instance, in both there are “Republicans” who are nearly at the “consistently liberal” outer edge, and “Democrats” nearly on the outer right. My guess is that this is based on party self-identification, but it would be good to know more.
