Friday, December 14, 2007

Difficulties Polling in Iowa

Here's something to think about the next time you see a poll predicting the results of the Iowa caucues. The New York Times has an article that discusses the difficulties of polling in Iowa.

All of the eccentricities of the Iowa caucuses make polling Iowa something most pollsters would rather not do. That night, it will be hard enough to depend on those entrance polls, taken moments before Iowans cast their votes. But in these last weeks ahead of the vote, to try and define the electorate and come up with a framework for polling is to rely as much on luck as savvy....

The biggest obstacle to polling in Iowa is drawing the sample in the first place. Some polls use lists of past caucus goers, but they risk missing first- time caucus goers, who accounted for more than half of the Democratic caucus goers in 2004, according to the network entrance poll. Other polls use lists of registered voters provided by the state, but they risk missing those who haven’t yet registered to vote but who plan to participate in the caucuses.

Turnout is so low, less than 10 percent, that few polls use random digit dialing techniques, like those used in national polls, to create a sample of caucus goers. On CBS News’s Web site, Kathy Frankovic, director of the CBS News Poll wrote: “Using random digit dialing techniques (in Iowa) could mean calling — and reaching— 10,000 households to interview only 500 attendees from each party.”

No comments: