rednecks for hillary

[Adam Griffin]

2012 White Vote by County
2016 White Vote by County

White Vote Swing by County, 2012-16

Note: the 2016 and 2012 white vote calculations were done by two different people (reagente & myself, respectively) and used different models to calculate the respective shares. If you see a county that looks completely out of step with its surroundings on the swing map, then it's likely an issue related to reconciling the two different models (this is most common in areas where the white vote is relatively tiny or the Latino population is relatively large; more room for error with even small adjustments in minority turnout/support). In some cases, my model is likely off while his is off in other cases. However, broadly speaking, our models were in general agreement in the vast, vast majority of counties.

Anyway, broadly speaking, Clinton improved over Obama the most throughout large portions of rural MS (likely due to lower turnout across the board), swathes of eastern NC and north/middle GA, and various rural coastal counties. In many of these places, black turnout dropped like a rock, which obscures the fact that Clinton did better among whites when merely looking at overall margins and comparing them to 2012.

In terms of her best rural areas in absolute terms: western NC, TN & AR. White rural Democratic vote share is heavily correlated with both latitude (even within the South itself) and race in the modern era (i.e. the fewer black voters, the more rural Southern whites tend to be).

[Adam Griffin]

2012 White Vote by County
2016 White Vote by County

White Vote Swing by County, 2012-16

Note: the 2016 and 2012 white vote calculations were done by two different people (reagente & myself, respectively) and used different models to calculate the respective shares. If you see a county that looks completely out of step with its surroundings on the swing map, then it's likely an issue related to reconciling the two different models (this is most common in areas where the white vote is relatively tiny or the Latino population is relatively large; more room for error with even small adjustments in minority turnout/support). In some cases, my model is likely off while his is off in other cases. However, broadly speaking, our models were in general agreement in the vast, vast majority of counties.

Anyway, broadly speaking, Clinton improved over Obama the most throughout large portions of rural MS (likely due to lower turnout across the board), swathes of eastern NC and north/middle GA, and various rural coastal counties. In many of these places, black turnout dropped like a rock, which obscures the fact that Clinton did better among whites when merely looking at overall margins and comparing them to 2012.

In terms of her best rural areas in absolute terms: western NC, TN & AR. White rural Democratic vote share is heavily correlated with both latitude (even within the South itself) and race in the modern era (i.e. the fewer black voters, the more rural Southern whites tend to be).

I grew up in PG County, Maryland. The result there does not surprise me.

It seems very inaccurate though, considering the county swung to Hillary.

Yeah, PG's relatively small white population means that even a small variance in projected white/black share of electorate in one or both of our models could throw the swing off considerably, hence the initial disclaimer. It's too bad GA is like the only state that collects and publishes racial turnout data all the way down to the precinct level, as that would make the projections even more accurate regardless of a county or precinct's racial makeup.