Can Iowa Caucuses be Polled?
John Zogby asks, “Can the Iowa caucuses be polled accurately?” and then more-or-less answers the question.
The premise is straightforward enough:
The Iowa caucuses require voters to go to a local school, church basement, private home or similar meeting place to spend between 90 minutes and two hours to register their preference. The process is a mixture of discussion, debating, a little horse-trading, and some consensus-building between neighbors. Anything can happen.
Not to mention that the Democrats and Republicans have very different processes for winnowing down the vote. Given, then, that intensity matters much more so than in standard primary contests and that people wind up voting for their second or third choice candidate after the initial balloting, isn’t standard polling meaningless?
Yes and no.
Zogby does a detailed analysis of his firm’s polling for the 2004 Democratic caucuses, which went in a span of ten days from a two-way race between frontrunner Howard Dean and Dick Gephardt to finishing off both men’s chances of getting the nomination.
When the final results of the caucuses were known, the four candidates finished in exactly the order indicated by the Zogby polls before the caucuses, though the percentages were different because undecided voters had finally made up their minds.
Can we predict the exact results of the Iowa caucuses ahead of time? The answer is simple: no. But that is not the purpose of political polling. As I mentioned earlier, there is no way to predict the neighbor-to-neighbor dynamic inside a caucus setting, and especially the effect that setting will have on those caucus-goers who show up to the events yet undecided.
But broad contours of the political landscape in Iowa can be determined by pre-caucus polling – the rest is up to the voters.
Of course, there’s also more than a bit of luck involved. Zogby’s final pre-caucus poll had John Kerry, Howard Dean, and John Edwards in a statistical dead heat (25-22-21). While he got the order right, he certainly could not predict it with any confidence whatsoever with those numbers.
So, yes, polling the Iowa caucuses provides some valuable insights into what the voters are thinking and what trends we’re likely to see. But, unless it’s a blowout, the polling isn’t particularly helpful in handicapping the horse race.