October 29, 2007
New IA Poll Tomorrow....But A Caution
Via Mark Halperin's The Page comes word that University of Iowa political scientist David Redlawsk will release a new Hawkeye Poll of likely caucus goers tomorrow. Just a reminder: The methodology of the previous U. of I. poll was unusual in a few respects: they asked an open-ended vote question (that asks respondents to volunteer their choice without prompting), they sampled from a directory of listed telephone households (rather than from registered voter lists or via random digit dial or voting list) and used a screen that was wider than other polls. More details in my post on their August poll.
We have not included the previous Hawkeye polls in our Iowa chart because of the use of the open-ended question. This is not a statement about the quality of the survey. As I wrote back in August:
This [open-ended question] undoubtedly provides a tougher test of voter commitment, but also produces a much bigger undecided and renders the results incomparable to other Iowa polls.
Of course, we do not know what methodology the Hawkeye poll used this time, but the news advisory does promises that the new results will show how candidate support "has changed since August." So presumably the methodology has been held constant.
By Mark Blumenthal on October 29, 2007 3:10 AM
Comments
The methodology is out on their website.
You may want to view the survey results and pay special attention to the sample dsigns used in the "Iowa Project 2000" located at
http://www.tufty.net/
For instance, the 03/04/99 survey on pages 2, 3, 5 and and the crosstabs for Q24.
Craig Tufty
Posted on October 29, 2007 1:54 PM