May 6th polls suggest split decision, with Clinton gaining some ground in NC

May 6th now appears to be Barack Obama's last hope to force Hillary Clinton out of the race before the end of primary season. The latest polls from Indiana and North Carolina suggest a split decision, a result that would be most damaging to Clinton's campaign though not enough to force her out of the race:

• Survey USA released a poll from Indiana showing Clinton in the lead, 52% to 43%. This is the same winning margin Clinton enjoyed at the beginning of April; two weeks ago, she led by 16%.
• The partisan gap is less dramatic than in some other Indiana polls we have seen, but Clinton polls much better among registered Democrats (+12%) while Obama narrowly leads among registered Republicans and independents. As expected, Obama's strongest regions are Indianapolis (where he leads by 9%) and Northern Indiana (bordering Illinois), but he only has a statistically insignificant 1% edge in the latter.

• In North Carolina, meanwhile, ARG finds Obama leading by 52% to 42%. This is essentially the same margin as last week, when Obama lead by 11%.
• Clinton keeps the race within single-digits among registered Democrats, 51% to 43%, while Obama trounces her 58% to 35% among independents.

• Update: Well, well, well: PPP just released its latest poll from North Carolina and, while Obama is still way ahead, there is some major movement in Clinton's direction. Obama leads 51% to 39%; last week, he was ahead by 25% so his lead has been cut by half since the Pennsylvania primary.
  
• PPP had been showing 20%+ Obama leads in North Carolina for a few weeks now, so last week's survey wasn't isolated. Note that PPP's turnout model appears to be favorable to Obama; PPP predicted a 4% victory for Barack in PA. Did they change their turnout model accordingly, or did Clinton pick up 13% within the original model?
  

In order to score a credible enough victory to be taken seriously in the final stretch (not the type of victory that would change the fundamentals of the race), Clinton does not need to win both states on May 6th but she cannot afford a significant loss in North Carolina; she probably has to keep the margin well within single-digits but polls confirm that will be a tall order for her. In Indiana, Clinton needs a comfortable victory of the size she enjoyed in Ohio and Pennsylvania, especially considering that the latter two were demographically better suited for her.

ARG, PPP and SUSA suggest that Clinton can meet such expectations; a mid-single digit North Carolina loss and a double-digit Indiana victory are within her grasp and a small shift in her favor in the next week would easily push her above the threshold in both states. Note, however, that Clinton's numbers in Indiana are among the most favorable to Clinton that are being recorded; some Indiana surveys have Obama leading (LA Times/Bloomberg, for instance).

Furthermore, Clinton benefited from a post-Pennsylvania bounce this week that is likely reflected in these numbers. Whether she can keep her momentum going, of course, is the question. For now, Gallup's latest tracking poll suggests that Clinton has managed to close the gap with Obama and challenged his claim to inevitability but she has not wrestled away an advantage. The two candidates have been in a statistical tie for three days now, with Obama regaining the smallest of edges today (47% to 46%; in the general election, Clinton is still leading McCain by 3% while Obama is tied with the Republican).