again: georgia is much closer than i expected. but, i still think trump has the edge.
why is it closer than expected? i know the narrative is about black and latino voters, but if you check the data you'll no doubt see that these demographics decreased their support for biden overall, not increased it. that's a longstanding trend. my analysis was that the polling error was a systemic bias that overestimated the effect of educated voters swinging away from trump in the south, and that seems to be broadly correct. i expect it to hold in arizona, in the end, too. so, why was georgia different enough that it almost swung?
when you're talking about small margins like this, you can essentially pick and choose. if trump wins by 800 votes in the end, you could assign that to anything - there may be 800 transgender voters in georgia, for all i know. there may be 800 executives for companies making steroids that compete directly with hydroxychloroquine. there may be 800 people in the georgia chapter of the ivana trump fan club. and, if anybody is going to defeat donald trump, it really ought to be fans of ivana trump.
and, i'm sure there would certainly be 800 white evangelical christian seniors that turned on him as a phony christofascist.
so, we can play this game in multiple different ways, but if we're to ask "what changed the most?", i suspect the right answer in the end will be that biden pulled enough christians to give trump a run, even if he didn't actually win. that doesn't make it a better or worse tactic than trying to win black voters if both tactics fail, in the end. however, it's worth pointing out which tactic involves swinging voters and which doesn't.
but, i'm surprised that it almost worked, if my intuition works out to be correct.
so, if biden pulls ahead by a few hundred votes, don't get too excited. you still have to wait until they're done.