Why the Iowa Straw Poll Maybe Doesn’t Matter All That Much

Aug 14 '11

Nate Silver has two quite interesting posts (“here”:http://fivethirtyeight.blogs.nytimes.com/2011/08/13/why-ames-actually-matters/ and “here”:http://fivethirtyeight.blogs.nytimes.com/2011/08/13/youve-just-won-the-ames-straw-poll-what-are-you-going-to-do-next/) making the case that the results of the Iowa straw poll contain a lot of useful information, in particular concerning who is going to win the Iowa caucuses the following year.

But just to play devil’s advocate, here’s a graph displaying how much predictive power the straw poll has had in the past. The poll has “been held five times”:http://en.wikipedia.org/wiki/Ames_Straw_Poll before, and one year there was a tie (1996, between Bob Dole and Phil Gramm). So we have six past winners – here’s how they did:

Bottom line: One winner of the Iowa Straw Poll has gone on to be elected president, George W. Bush. Two of the six got the Republican nomination. This is what it is: a snapshot of how some dedicated activists in the Iowa Republican Party feel about the presidential candidates about half a year before the Iowa Caucuses. As such, it is likely to be a decent predictor of how dedicated activists in the Iowa Republican Party feel about the presidential candidates at the time of the Iowa Caucuses, but far from a perfect one. The best way to think about this is probably as yet one of many early opinion polls, and as research by “Christopher Wlezien”:http://www.temple.edu/polsci/wlezien/index.htm and “Robert S. Erikson”:http://www.columbia.edu/~rse14/ has repeatedly shown (see “here”:http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=257300&fulltextType=RA&fileId=S1049096504045068 and “here”:http://ijpor.oxfordjournals.org/content/19/1/74.full for example) “early polls only tell us little about the final election outcome”.

Now that being said, it is clear that the Iowa Straw Poll can drive behavior in the future. For example, I will go out a limb and say on the basis of the straw poll, “Tim Pawlenty will not be elected president in 2012”:http://thecaucus.blogs.nytimes.com/2011/08/14/pawlenty-drops-out-of-republican-race/. But then again, I probably would have said the same thing two days ago on the basis of results from traditional polls as well….


Update: Great “post with much more sophisticated analysis”:http://pollsandvotes.com/PaV/?p=139 from “Polls and Votes”:http://pollsandvotes.com/PaV/?p=139 that basically makes the same point more convincingly!