The news lately has been all about Barack Obama's recent gains in Iowa polling, putting him in a statistical tie with Hillary Clinton (and leading her in raw numbers in at least two polls), and leaving John Edwards a close, but disappointing, third. Obama's move is real enough, at least in the polling, but don't count Edwards out just yet. Noam Scheiber of The New Republic explains:

I agree with E.J. Dionne--I think Edwards takes second in Iowa. My thinking is this: Clinton and Obama are engaged in a death-match there. Someone is going to win that death-match, and someone is going to lose it, and the person who loses it is going to be in big trouble. That's because the person who loses will not only have lost on semi-substantive grounds (by which I mean not just health-care or foreign policy but whether they have the experience, judgment, character, etc. to be president), but because they will have been diminished in the process. (The only thing worse than a nasty, ruthless pol is a nasty, ruthless loser.) Edwards, by contrast, seems to have toned down his rhetoric a bit, content to let the two titans fight it out. On top of that, given his support in the state in 2004, I think he starts with the highest floor of any of the three candidates.

Read it all at the link above. The above reasoning is sound as far as it goes, and Edwards adviser Joe Trippi should know all about the "murder-suicide" scenario in Iowa - he was Howard Dean's maven as Dean and Gephardt turned their close-favorite campaigns into third and fourth places in 2004 with a mutually withering negative campaign. Hillary and Obama both witnessed that, too, so they understand the risks. Still, negative attacks and responses can escalate quickly in Presidential politics, as evidenced by the recent Giuliani-Romney feuding on the Republican side.

But Edwards has other trumps as well.

He has been organizing in Iowa since at least late 2002, and won 18% of caucus delegates in 2004. His support also comes heavily from blue-collar workers generally, and older Democrats, which demographic groups tend to show up at caucuses in disproportionate relation to their share of the population. Hillary Clinton's vote will also come predominately from these groups, but Barack Obama wins the lion's share of the younger voters and those who have never before attended a caucus.

Primary polling is difficult because participation depends on many factors, so how pollsters weight for "likely voters" directly affects the accuracy of their predictive value. Caucuses are even tougher, because it involves more commitment than any primary. It's more than just showing up anytime between 7 a.m. and 7 p.m. and hoping the line isn't too long; caucusers must show up to a usually poorly-heated location at 7 p.m. on a week night (when there is likely to be a good college football game on television) and be prepared to spend three or more hours taking votes and realigning with others. In early January, snow or ice on the roads is not unlikely, and bitter cold is almost assured. You can't mail in an absentee caucus vote.

A key factor in Iowa's caucus is that in order to win any delegates at a precinct, a candidate must have at least 15% support. Those that don't must join another candidate (or "uncommitted") who did make the cut. This process repeats until all the groups constitute at least 15% of those attending. It's very important, therefore, not only to be the first choice of as many voters as possible, but also to be the second choice. If Richardson or Biden supporters fail to get to 15%, as seems likely at most precincts, they must join another caucus. If that is Edwards, it's a big boost for him. Also, in those few precincts where Hillary or Obama voters fail to muster enough support, their voters won't go to the other, so they are up for grabs.

Obama's young supporters may pass the pollsters' criteria for "likely" attendees, but there is a long record which says they will not show up in the numbers predicted. Also, he is strongest in areas of the state where Democrats haven't always run so well. That would be a huge advantage in a general election, but Iowa Democrats award delegates to precincts based upon their past Democratic vote - a smaller precinct in a heavily Democratic area often has more delegates to bestow than a larger precinct in a competitive area or a Republican stronghold.

It doesn't take much imagination to see Edwards pulling out second place, or even first, given the peculiarities of the Iowa system.