A couple of random thoughts heading into the May 6 primaries:

1.) There are four polls out today that previously came out last week. All four show Obama gaining in Indiana and Clinton gaining in North Carolina. In Indiana, Clinton's lead is down from 8 to 5 (PPP), from 9 to 8 (ARG), 7 to 4 (Insider Advantage), and from tied to down 2 (Zogby). In North Carolina, Obama's lead is down from 12 to 10 (PPP), from 11 to 8 (ARG), 5 to 3 (Insider Advantage--actually more like 5 to just under 4), and 16 to 8 (Zogby). I don't know precisely where that leaves us, but, qualitatively, I'd guess we're looking at a "solid, but not as big as it could have been" win for Hillary in Indiana, and a "closer-than-expected, but not super-close" win for Obama in North Carolina.

2.) Demographically, the two groups to follow are African-Americans (obviously) and college-educated voters. Mark Blumenthal puts the likely black vote-share at 9-12 percent in Indiana, and high 20s to high 30s in North Carolina. The bottom end of those ranges should be good news for Hillary, the top end should be good for Obama. Except! We don't yet know how African-Americans are reacting to Wright. I speculated last week that some black voters could be even more pessimistic than white voters about what Wright's done to Obama's electability. If Obama's margins among black voters slip, that could be the reason, and things could get dicey for him regardless of turnout. (Interestingly, at least in Indiana, the robo-polls show Hillary doing slightly better among African Americans than the live-interview polls. I wonder if they're picking up on a queasiness black voters are less comfortable sharing with human interviewers.)

As for college grads, this goes back to a debate John Judis and I had after Pennsylvania. Obama rarely loses this demographic, but he lost it last time out, and losing it again would be worrying--a sign that his coalition is shrinking. On the other hand, per the back-and-forth Mike and I had last week, if there were ever a demographic that might respond well to Obama's attack on gas-tax pandering (and badly to Hillary's attack on experts), it's college grads. The gas tax debate could help him nail down this group.

3.) Is it possible that Obama has a bigger cushion than we realize in North Carolina thanks to early voting? Blumenthal cites an analysis by a George Mason professor suggesting that roughly 400,000 people have already voted. The Charlotte Observer puts the upper limit of tomorrow's turnout at around 1.5 million, meaning more than a quarter of the vote may have been cast. If true, that could be a lifeline for Obama. The polling on early voters is sparse, but suggests they lean his way by close to a 60-40 margin.

Update: Via Halperin, this Boston Globe item says nearly 500,000 people have already voted in North Carolina. I guess the question is how many of them voted before last Monday, when Wright let it rip...

Update II: This just-released SurveyUSA poll puts early voters at about one-quarter of the total in North Carolina. The early voters prefer Obama by a 57-41 margin. For what it's worth, Indiana also does early voting, but it's a much, much smaller share of the likely electorate. (Only about 3 percent according to SurveyUSA.)

Update III: A final Zogby poll out today show's some widening for Obama in NC, for what it's worth. The final Zogby in Indiana is basically unchanged, as is the final Insider Advantage poll in NC.

--Noam Scheiber