Looking at the current polling situation in Iowa, it appears that Edwards is close, but still behind. Pollster.com currently projects Edwards at 23.4%, but Clinton at 27.2% and Obama at 29.0%. Real Clear Politics shows a six-poll simple mean with Edwards at 23.2%, but with Obama and Clinton tied at 28.3%. So, Edwards is clearly in the mix, but he trails both Clinton and Obama by anywhere from 4-6% with only ten days to go.
Or does he trail Clinton and Obama? There are at least four theories floating about that argue Edwards is either tied in Iowa, or that he is actually ahead. In the extended entry, I consider all four of these theories. Spoiler alert: I think that the strange caucus system means that Edwards does have more support in Iowa than polls are showing, but right now it isn't enough to win. He needs to close the gap in public polls, too.
The Edwards Field Advantage Theory. The Edwards campaign claims to have precinct captains in over 90% of all Iowa precincts, many of them manned by people who caucused for Edwards in 2004. This advantage is born out of Edwards spending more time in Iowa than any other candidate, some arguing that he has been continuously campaigning in the state for nearly six years now. Certainly, it is an impressive figure, especially when one considers that the Obama and Clinton campaigns have not released comparably figures. However, a common misconception about strong field operations is that such operations allow a campaign to defy polling, and dig up a few "hidden" percentage points. The truth is that likely voter polls, when properly conducted, take all aspects of a campaign into account, including field. Field operations, like both paid media and free media communications operations, are simply another form of voter contact designed to make potential voters more likely to go to the polls and select a given candidate. The vast majority of field contacts occur before Election Day, and as such are actually measured in polls (or, at least they are measured in polls that accurate gauge how likely people are to vote). Sure, there are field contacts made on Election Day itself, but there are also paid media and free media contacts made on Election Day, too. Thus, unless polling for a given election sucks, no campaign should be able to shock the world simply due to an excellent field campaign. Conclusion: Don't Buy This
The Edwards Likely Voters Theory. This theory picks up where the first theory falls flat. If public polling is accurately determining likely voters, and thus there is no hidden field advantage for the Edwards campaign in Iowa, doesn't Edwards actually hold an advantage among likely voters? A common theme is that Obama in particular is relying on younger voters who have never previously participated in a caucus, and who might not even be around on January 3rd. The problem here is that among the last three polls to release numbers for both less likely and more likely caucus goers, Edwards shows no improvement among the more likely caucus goers. According to Rasmussen, Edwards is as 22% in both sets of voter. According to Insider Advantage, Edwards drops from 30% to 26% when less likely caucus-goers are removed from the equation. The ABC News / WaPo poll showed a wide range of possibilities among likely voters, with no clear likely voter advantage for any one candidate. A little further back, the early December Newsweek poll also showed Edwards doing worse among more likely voters than among less likely voters. So, unless the likely voter models of these polls all completely suck, this theory doesn't pan out either. Conclusion: Don't Buy This
The Edwards Second-Choice Advantage They Now we start entering the realm of more plausible theories. As everyone reading this blog knows, second-choices matter in the Iowa caucuses because of the 15% (and up!) viability threshold for delegates in Iowa precincts. This should translate into a boost for Edwards, since every single poll out of Iowa shows Edwards leading among second-place choices. The one question mark is just how far ahead Edwards is among second place choices. His advantage in this category varies significantly across polls, but usually falls in the high-single digits. It is difficult to say how much of an advantage this will translate into on caucus night. Certainly, it will be worth at least one point, but it is also unlikely to be worth more than two points. Conclusion: This seems likely to be hiding a 1-2% Edwards overall boost, and a 1% boost relative to both Clinton and Obama.
The Edwards Rural Advantage. Looking back at 2004 Iowa results by county, it certainly seems that Edwards did much better in the less populated areas of the state. Now, I haven't found any polls confirming that Edwards continues to lead among rural voters. Instead of providing such crosstabs, most polls instead ask really empty and stupid questions like "who is the strongest leader," and "who has new ideas" that serve no purpose apart from increasing the vacuity of the national election narrative (I bet that many poll respondents think such questions are matching quizzes to the national narrative more than anything else). However, there is no reason to assume that anything has changed this time around, especially when national polls confirm that Edwards does better in rural areas than in more densely populated regions. Certainly, along with the Kucinich deal and being a favorite second-choice pick, this was a also factor in his dramatic six-point improvement from the entrance poll in 2004. It will probably help him again this time. Conclusion: This is probably hiding 1-3% for Edwards
So, overall, I think there is cause to suspect that Edwards will do at least 2% better, and possibly 5% better, from the entrance poll to the final results this time. Will it be enough for him to close the deficit he faces to Obama and Clinton? Right now, my best guess is that he would have a decent chance to pass Clinton, but his second-choice, rural vote cocktail would not be enough to pass Obama right now. The reason for this is that Obama does a little better among second-choice voters than Clinton, seems to have a slightly more favorable trend than Clinton, appears to do a little better among the most likely voters than Clinton, and also appears to be very slightly ahead of Clinton overall right now. So, if there was a Christmas Eve Iowa caucus, I would predict a three to five point Obama victory, with Clinton zero to two points ahead of Edwards for second. Perhaps Edwards would even edge Clinton by one point.
Anyway, As I wrote at the beginning of the piece, if Edwards is going to win Iowa, it will be necessary to see him closing the gap in polls, not just in hidden strengths. He should see a hidden improvement from the entrance poll, but right now I don't think it will be enough to win.
