Problem with option B is that rural white deep south men are not that big a demographic.
Deep South - about 12% of the population
Rural Deep South - about 6%
White rural Deep South - about 4%
White male rural Deep South - about 2%
... who voted for Clinton - about 0.2%
How many Lesbians are there living in the US?
If you consider 5% of the population gay and half of the population female, about 2.5%?
And how many of them might have chosen Trump over Clinton? And why?
I assume that the 5% figure also includes Bisexuals. The number of Lesbians would thus be closer to 1-1.5%.
In either case, to match the number of Stein voters, Trump would have had to definitionally won a majority of Lesbians, which clearly didn't happen.
B is clearly the largest out of the options provided in the OP (it isn't even close). The only question is if A or C is bigger.
Not really, let's go with the 5% estimate of the LGBT population share. Lesbians are probably at most 2%. And Trump did horribly with LGBT people, exit polls say he had 14% of the vote. Let's just assume lesbians voted like LGBTs at large (though I have a hunch that Donald would have even less support among lesbians). Then the share of the voters is 0.28%, far less than Stein's 1.1%.
Now using the broadest definition of Deep South - Louisiana, Mississippi Alabama, Georgia, South Carolina, northern Florida, western Tennessee, and eastern Texas...
Deep South - 13.4%
+ male and heterosexual (~46%) - 6.2%
+ white non-Hispanic (~55%) - 3.4%
+ rural (~34%) - 1.2%
+ voted for Clinton (~20%? probably generous) - 0.24%
No question Stein voters are the biggest group. The real question is whether there were more As or Bs. But my estimate for the lesbian share was kinda generous, so I would say
C >>>> B > A
Poll
