I'm not too surprised when I see Herman Cain doing better than Rick Perry in recent polling, but I was surprised to see Ron Paul beating Rick Perry in NBC's new poll of Iowa. So, I started getting ready to post... and found out there's more than meets the eye here.
Caucuses are apparently hard to poll.
The facts: NBC News poll by Marist, 3913 adults, MoE 1.7. Likely caucus goers pare it down to 371, MoE 5.1. Mobile and Landline handling.
Mitt Romney leads the poll at 23, Herman Cain follows close at 20, Ron Paul sits in third at 11, and Rick Perry falls below the Pauldoza line to fourth at 10. The Cain moment continues.
But here's what struck me: the prediction of Caucus turnout. In 2008, both parties set records for the Iowa caucuses. Democrats nabbed 12% of all registered Iowa voters, while Republicans drew 6.2%, bringing total two party turnout to 18% of the registered electorate in Iowa.
NBC's poll projects new records to be set in 2012. Of the 2836 designated registered voters in their poll, 371 of them are designated likely GOP caucus goers, 13%. That's more than the Democrats got in the crest of their wave year, 2008, with a bruising two way primary fight. That's more than double the Republican record turnout.
Is this possible? Of course it's possible. Is it likely? I'm skeptical, so I begin to be skeptical of whether polling geared to predict elections is capable of predicting caucuses well at all.