Four opinion polls tracking the Iowa caucus vote came out yesterday. All of them were either taken on December 28 exclusively or ended on December 28. The news reports are saying they are all over the place, which is true. Who should you trust?
Well, you should trust Matty Boy, professional mathematician.
I have put the polls from left to right, ARG, Insider Advantage, NBC News/Marist and Rasmussen. It just so happens that's alphabetical order, but more importantly the order goes from the highest number of "none of the above" to lowest, marked by the black line. My assumption is that less than one week out from a caucus, that number shouldn't be very high anymore. Caucusing in Iowa is a real commitment of time, and there shouldn't be 10% or more of the people planning to attend still on the fence.
So, tip #1: Less undecided equals more reliable.
Notice whose numbers rise as undecided fall, Paul and Santorum. To a lesser extent, Perry and Romney do better as more people make up their minds.
Notice whose numbers tank as the undecided fall, Gingrich and Bachmann. Earlier in the week, a Gingrich spokesman said they would be happy finishing fifth in Iowa. I thought that was a mighty low hurdle, but his internal numbers must have been looking like these.
Huntsman is making no effort at all in Iowa. Does that make sense? The caucuses produce ZERO delegates to the national convention, just delegates to a state convention. Yeah, Huntsman is probably shepherding his limited resources wisely.
I'll wait until Sunday to make a prediction for Tuesday, but my guess right now is Romney and Paul jockeying for first, Santorum alone in third, Perry over Gingrich for fourth and Bachmann trailing badly in sixth. Some jackass is putting out radio ads asking people to caucus for Palin, though her money is not connected to this turkey. I seriously doubt the Palin "grassroots" will even catch Bachmann, and will likely not do as well as Huntsman. The folks in Iowa like being fussed over.