Readers of John B. Judis’s article “Poll Potheads” (January 11, 2008) should be aware of the following regarding Mr. Judis's assertions: First, I never wrote in my column in The New York Times anything that would support this characterization: "Could it be that voters lied to pollsters this time, too? Andrew Kohut, a pollster for the Pew Research Center, thinks so." What I concluded in the NYT was: "In New Hampshire, the ballots are still warm, so it's hard to pinpoint the exact cause for the primary poll flop. But given the dearth of obvious explanations, serious consideration has to be given to the difficulties that race and class present to survey methodology."

Second, Mr. Judis presents very incomplete data to refute the possibility that the problem could be that lower socio-economic status refusers of surveys may have different opinions than those pollsters interview. He concludes the following based on looking at one cross-tab from a UNH poll: "Where Clinton dramatically picked up support from the pre-election poll to the final poll was among voters with college degrees and higher. That's exactly the opposite of what Kohut's version of the Bradley effect would predict." But a comparison of more comprehensive Gallup pre-election survey results with exit poll findings show that the gaps were greatest for poor people and less well-educated people. By income, Gallup's pre-election Obama margin was 22  percent greater than the exit poll margin among voters earning less than $50,000 per annum, compared to just 11 percent for those earning more than $50,000. By education, Gallup's pre-election Obama margin was 13 percent greater than the exit poll margin among voters with high school education or less, compared to just 7 percent among those who post graduate educations. In other words, the overstatement of the Obama margin in Gallup's pre-election poll was greatest among the kinds of people who might be different than those of the same socio-economic class who pollsters often fail to interview. Mr Judis's reporting of what I said was wrong, and the data he uses to disapprove the possibility of the point I raised are very weak and incomplete.

Andrew Kohut, President

Pew Research Center




By Andrew Kohut