1: Analysts who number-crunch early votes are seeing African-American turnout down in Florida and North Carolina. If you're a Democrat, that is not good. In Florida at least Hispanic turnout is higher. A decline in African-American turnout could be one key reason why Florida is looking closer than it seemed a week or two ago. Of course, there's still a week of voting left. So the African-American turnout number could change. In North Carolina at least that 'underperformance' of the African-American vote is likely significantly driven by new voter restrictions - most importantly, restricting the amount of time in which all early voting stations are open. Those GOP led voter restrictions appear to be having an effect. What's not clear yet is whether they're just pushing the same amount of early voting (and African-American voting) into fewer days or reducing it overall.
2. One dynamic of this race is Trump's disproportionate share of the vote from non-college educated white voters. The key is that that slice of the population is relatively over-represented in the big swing states. That's the main reason why Clinton is slightly underperforming Obama in most swing states and relatively over-performing in a number of reddish states which are younger and less white. So, for instance, Arizona is very close, even though Clinton likely won't win there. Texas is closer than it usually is even though she almost certainly won't win Texas. What it amounts to is a net negative since a lot of that additional support in red states - mainly college-educated whites who voted for Romney but won't support Trump - is, in hard electoral terms, wasted.
3. One fascinating part of the 2016 cycle: In part because of dynamics I noted in Item #2 there is a decent chance that Donald Trump will win both of the archetypical battleground/swing states of the aughts (Florida and Ohio) and yet still lose the election. That's because of states like Virginia, Colorado and North Carolina. Virginia and North Carolina weren't even considered swing states in the Bush years. Now Virginia and Colorado will also certainly go to Clinton and there are better than 50-50 odds North Carolina will too.
4. Polls are weird. Even though the polls have been all over the place, my main concern over the last week hasn't been the presidential result but the senate. (See The TPM Senate Scoreboard).Republicans are quickly lining up behind a pledge to continue the Supreme Court blockade which began with Antonin Scalia's death through the entirety of Hillary Clinton's term if she becomes president. So getting at least 50 senate seats (a tie broken by a Vice President Kaine) is truly critical. I'm still pretty confident of the presidential outcome. But while the race at the top of the ticket is tightening, polls over the last couple days seem better in the Senate. Pat Twomey had been hanging on remarkably well for months, given that he's significantly more conservative than his state and is running in a state where Clinton will likely win with a decent margin. In the last week or so, though, things seem finally to be moving clearly against him. I always figured that even if he had a one or two point lead on election day he'd likely lose because of the downdraft from the top of the ticket. As of this morning, he's 4.9 percentage points behind.