Do Open-Ended Questions Measure "Salient" Issues?

Closed-ended questions dominate most interview schedules. Yet the almost exclusive use of this form did not arise because open-ended questions, its major competitor, proved to be weak indicators of public opinion. Instead, responses from open-ended questions proved more difficult and expensive to code and analyze than those from closed-ended questions. Although such practical concerns are important, the real task of survey researchers is to measure public opinion accurately. Using an experimental design, this article tests whether open-ended questions measure the important concerns of respondents—one of the long-claimed advantages of this format. The results, on balance, show that open-ended comments reflect such concerns, suggesting that pollsters may want to include more of these questions in their surveys of public opinion.