GA-Atlanta Journal-Constitution: Trump +2

[Adam Griffin]

[Adam Griffin]

Crosstabs?

[Adam Griffin]

One problem I noticed here: their racial crosstabs are simply "black" (32%) and "white" (68%).  Whites were 61% in 2012 as defined by SoS and while I'm sure some of those people who were only categorized as "other" or "unknown" by SoS are in fact white, they're basically ignoring all of the non-black, non-white vote here. It's not huge by itself in sheer vote totals but when you're talking about even a small group (say, 5%) swinging from 80% R to 70% D, it matters. It's enough to take the race from Trump +2 to Clinton +3.

The black share is about one point too high in my opinion, but the fact that the other 5% or so of the electorate (that's not white neither in reality nor by SoS categorization) is missing more than cancels it out. Would still be Clinton +1.

[Adam Griffin]

That 5% undecided will go at least 4-1 Trump. As long as Trump has a nominal lead in the GA polls, it's very hard to see him losing the state.

If you look at when undecideds have broke toward the GOP in the past (how many days before the election), then we're entering that period right now. If it happens, we're mere days away from seeing polls that would suggest Trump will win Georgia by as much (or more) as Romney won by in 2012.

[Adam Griffin]

She doesn't necessarily need 27%. She can get by with as little as 23% potentially; even less if Johnson actually does well. As I always say:

So if we assume that Clinton's white share of the vote (23%) remains identical/none of the "other partiers" break for her, females reinforce the black vote at Obama 2012 levels of support and the "other" racial groups go for Clinton by the same margin as in the poll (Clinton +22)...

Group	% of Voters	Dem Support	% of Vote
White	59	0.23	13.57
Black	31	0.95	29.45
Other	10	0.61	6.1
Total	100		49.12

That's victory if Johnson takes 1.9 percent of the vote or more.

Idealistically, I know that if Clinton carried the same percentage of the vote across each race that the President did in 2008 in Georgia, then she'd win 50-51% of the vote from the likely 2016 electorate.