That Newsweek poll Isaac cited looks intriguing, but it doesn't actually tell us much. The margin of error for the likely-voter portion of the poll, where Obama leads Clinton by 4 points, is plus or minus 7--i.e., Obama's lead is statistically meaningless. In fact, Obama and Edwards, the third-place candidate among likely voters, are only separated by 6 points, which means they're in a statistical tie, too. Correction: Brad Plumer points out that I've misinterpreted the margin of error here. The +/- 7 percentage points only applies to a particular candidate. Since Obama is polling at 24 percent among likely voters, we know with some high degree of confidence that Obama's actual level of support is between 17 and 31 percent. That obviously isn't so helpful for figuring out his absolute level of support. But, then, we don't care so much about his absolute level of support, just his support relative to Hillary's. The formula for figuring that out is actually pretty complicated--and relies on information, like the sample size of likely Democratic voters, that we don't have. (See Kevin Drum on this here.) That said, there's a pretty good chance Obama has some sort of lead over Hillary, even if focusing on his 4-point lead is misleading with such a high margin of error.* *If there's a pollster or math professor out there who can provide the formula for determining the probability that Obama has a lead over Clinton and Edwards in a poll of eight candidates, I'll try to get the sample size so we can figure this out. Update: For what it's worth, it occurs to me that we can basically back out the sample size, as long as we know the "confidence level" Newsweek was polling at... --Noam Scheiber