Mississippi whites who vote Democratic

[Del Tachi]

Pretty obvious.  Some college town people, some government employees here and there, Millennials (specifically in urban areas), your run-of-the-mill urban White liberals and a few Yellow Dogs who have somehow managed to avoid dying yet.

All the old Yellow Dogs still living are Democrats who vote Republican, not Democratic.

Not necessarily.  Jim Hood and Brandon Presley routinely win a majority of White voters in the more rural, conservative northern half of the state.  North Mississippi is the ancestral home of these "Blue Dogs".  This is the same region that went for Kennedy in 1960, gave Goldwater his lowest margins anywhere in the state, and went strongly for Carter in 1976 and 1980.

[Del Tachi]

Pretty much the same groups that vote Democratic in other Deep South states: there are just fewer of them due to no major metro areas or large cities (and fewer small cities/areas of density due to larger/fewer counties than a state like GA).

That's of course the case because Mississippi's economy is one of the weakest nationally, which is caused in part by a relative lack of educational attainment (which, likewise, drives down white rural support lower than in places like AL & GA).

Educational attainment for rural MS whites shouldn't be any worse than for rural Whites in AL/GA.  The fact that the uber-White Mississippi counties stick out on that map compared to AL/GA has something to do with whatever racial turnout dynamics you're using.

[Del Tachi]

Derived from my White vote by county project

Why do some almost-monolithically Black counties (Noxubee, Holmes, Claiborne) seem to have an unusually-high % of White Democrats?  I think it has something to do with how you're modelling Black turnout/partisanship.  What sort of assumptions are being made there?

[Del Tachi]

Thanks for the great posts, Adam and Reagente

There's evidence that Mississippi began losing raw White population before 2010, and it's definitely been the case since 2013.  If this effect is especially pronounced in majority-Black rural areas (which I suspect that it would be, just because economic prospects in these areas are so grim) then these counties' electorates might be notably less White in 2016 than they were in 2010.  This would elevate the White Dem % in these areas.  Waiting for 2020 Census numbers to validate this hunch.