The Morals Choice by Howard Shuman

The 2004 exit poll, carried out by the National Election Pool (NEP), asked voters to choose from among seven issues the one that mattered most in deciding their vote for president. "Moral values" came in first by a nose, chosen by 22 percent of the voters, followed by economy/jobs (20 percent), terrorism (19 percent), Iraq (15 percent), health care (8 percent), taxes (5 percent), and education (4 percent). The first place rank of "moral values" is seen by some as an important source of George W. Bush's increased vote margin in 2004. Others consider the result more a fluke in question wording than a finding. Neither view is correct, based on what we know from many years of research on survey questions.

First, open questions that require spontaneous responses, rather than choice among alternatives, are regarded by some as the gold standard for judging the importance of issues. This view is not persuasive. For example, in an experiment published in a 1987 Science magazine article by Jacqueline Scott and myself, when Americans were asked to name especially important events of the past half century, they seldom mentioned the "invention of the computer"; but when that alternative was listed along with the most frequent open responses (World War II, for instance), the computer came out on top! Quite likely the invention of the computer had been outside the frame of reference for the open question, but was judged to be highly important once legitimated by being offered as an alternative.

Furthermore, open questions probably suffer more than closed questions from salience effects. Forced to come up with a response on their own, people are apt to be influenced by what they have heard most recently on television or other news sources, and Iraq and terrorism have been much more in the news than the summary phrase "moral values." A closed set of choices can offer a more even playing field, provided it does, indeed, include all the main issues that respondents may wish to consider.

On the critical side in evaluating the NEP question, it is true that the percentages resulting from a list of alternatives are always a function of the number of choices and the way they are worded. Results from a single question can never indicate which issue is "most" important, and, of course, 22 percent is only slightly above the percentages shown for several other choices in the NEP exit poll. Identifying "the most important issue" is a matter of judgment and should be based on considering the findings from a number of questions. There is no quantitative certainty possible as to the conclusion, and there may well not be any such single issue.

More useful with this exit poll question, as with all survey questions, is to focus on differences between key groups. Based on the exit poll data, Bush voters were much more likely to choose moral values than Kerry voters: 35 percent to 8 percent. Moreover, the Los Angeles Times exit poll reported a similar result: Bush supporters were more than twice as likely as Kerry supporters to choose "moral/ethical values" from among twelve alternatives. These differences represent real findings, worth emphasizing. Whatever "moral values" means to Bush voters, they saw them as a great deal more motivating than did Kerry voters.

One further objection to including "moral values" as an alternative in the closed question is that it does not constitute a "concrete" or "specific" issue parallel to, say, Iraq or taxes. Whether or not that is the case-and the single word "Iraq" is also unspecific-the goal in framing a survey question is ordinarily to capture words that are meaningful to people, whether or not all the words are at exactly the same level of abstraction. To the extent that "moral values" is meaningful to a substantial number of Americans in deciding how to vote, it is a legitimate alternative.

Of course, it would be informative if respondents had been probed to find out what they believed "moral values" to stand for, but the self-administered exit poll did not do that. We do know from the Los Angeles Times exit poll that the alternative "moral/ethical values" was not only the most frequently chosen at 40 percent, but that one of the other alternatives was "social issues such as abortion and gay marriage." The latter was selected by 15 percent of the sample, and almost equally by Bush and Kerry voters. Evidently "moral/ethical values" is regarded as broader and more inclusive than such specific issues in its appeal to Bush voters.

Yet even if we conclude that "moral values" was an appropriate alternative to include in a question about the reasons for Bush's support, and that it reflects a significant aspect of his support, this does not show that it contributed to the increase in his popular vote margin over 2000. Other survey results suggest that issues falling under the rubric of "moral values" were about equally important for Bush's support in the two elections, though we lack additional measures (including intensity of feelings) that make it possible to estimate any change in impact between the two elections.

Moral values may have been highly important to Bush's election in 2004, but not more so than in 2000. Furthermore, the alternative "terrorism," which was of course not a possible option in 2000, showed an even larger difference in ratios than moral values between Bush and Kerry voters in 2004 (32 percent to 5 percent in the NEP exit poll). It probably made most of the difference in vote preference between the two years.

Howard Schuman is a professor and research scientist emeritus, University of Michigan. Most of his research has been done through the university's Survey Research Center. He is the senior author of Questions & Answers in Attitude Surveys: Experiments on Question Form, Wording, and Context (1981;1996) and of many other articles on survey questions.

Readers who wish to respond to this or other articles appearing in Public Opinion Pros, or to contribute commentary of their own in 800 words or less, should consult our author submission guidelines and editorial policies under "Letters to the editor and op-ed articles."