Exit Poll Spin: Buyer Beware

By George Bishop

Beware of false pundits bearing exit polls, claiming to tell us why George W. Bush won the 2004 presidential election. They don't know why Bush won, at least in part because the voters themselves can't tell us why they voted the way they did.

The initial postelection analyses told us that George Bush won because a lot of exit poll respondents picked "moral values" as the one issue that mattered most in deciding how they voted for president. Among those who said "moral values" was the decisive factor, 80 percent said they voted for Mr. Bush. In the battleground state of Ohio the exit poll results were even more lopsided: 85 percent of the "moral values" voters supported President Bush.

Now the pollsters have weighed in with a new twist, arguing that this explanation is misleading because the term "moral values" is an ambiguous catch-all that could mean different things to different people: abortion, gay marriage, stem-cell research, sex on television, and God knows what else. If only this poorly worded option had been excluded, they tell us, the results of the exit poll and the meaning of the election would have been much clearer.

But the other choices in the now-infamous exit poll question may also have meant different things to different people. Did Bush supporters, for example, have the same thing in mind as Kerry supporters when they selected the "economy/jobs" as the critical issue in their vote decision? Did both camps interpret the meaning of "Iraq" and "terrorism" in the same way? "Health care?" "Education?" "Taxes?" Not too likely. These are all more or less ambiguous Rorschach blots.

Table 1
Responses to "Moral Values" Question, by Vote for President

Question: Which ONE issue mattered most in deciding how you voted for president?
(Check only one)

Other polling gurus have also told us that the real problem was that the presentation of "moral values" as an explicit choice in a closed-ended format greatly exaggerated its apparent importance to the electorate. If only the question had been asked in an open-ended form, they say, all this interpretive confusion would not have occurred. As Howard Schuman reminded us, however, in his recent op-ed column in Public Opinion Pros, open-ended questions with seemingly "spontaneous responses" suffer from their own defects.