Response Alternatives

Response alternatives are the choices that are provided to respondents in a survey when they are asked a question. Response alternatives are generally associated with closed-ended items, although open-ended items may provide a limited number of such choices.

Response alternatives can take a number of different forms, related to the type of question presented. In a Likert-type item, in which respondents are asked the extent to which they agree or disagree with a statement, the response alternatives might be strongly approve, approve, neither approve nor disapprove, [Page 750]disapprove, and strongly disapprove. For a rating scale—for example, the rating of the job the president is doing—the response alternatives might be excellent, good, fair, poor, and very poor. Another example of a rating scale would be a “feeling thermometer” in which respondents would be asked to rate some individual or object on a scale from 0° to 100°, in which the degrees would represent the response categories.

Types of response alternatives may vary depending on the method used for data collection. In face-to-face and self-administered surveys, respondents can be presented with visual materials displaying the response alternatives. For example, in a face-to-face survey, a respondent can be presented a show card describing the various alternatives, and a similar description of the response options can be provided in a mail or other self-administered survey and for surveys conducted via the Internet. As a result, a larger number of response alternatives can be presented in these modes of data collection than in telephone surveys. The number of response alternatives that an average respondent can remember over the telephone is generally limited to five. If a larger number of response alternatives is desired, the question is typically divided into a root item, which is then “unfolded” into this larger number of choices. For example, if researchers were interested in a person's ideology along a liberal-conservative scale, they might first ask, In politics, do you generally consider yourself liberal, moderate, or conservative? Those who said “liberal” would then be asked, Would you say you are extremely liberal or somewhat liberal? and, similarly, those who said “conservative” would be asked, Would you say you are extremely conservative or somewhat conservative? Respondents who answered “moderate” would be asked, Do you lean toward the liberal side or lean toward the conservative side? The result would be seven response alternatives: (1) extremely liberal, (2) somewhat liberal, (3) leans toward liberal, (4) moderate—leans toward neither, (5) leans toward conservative, (6) somewhat conservative, and (7) extremely conservative.

Another consideration in presenting response alternatives is whether to provide choices such as “don't know” and “refused” as explicit options to respondents. Typically such choices are not read to respondents in face-to-face or telephone interviews and are not included as options in self-administered or Internet surveys. There are situations, however, in which “don't know” or “no opinion” responses are important to the researcher; in such cases an explicit “no opinion” response would be provided. For example, in the question, Do you think state spending on roads and highways should be increased, kept about the same as it is now, decreased, or don't you have an opinion on this issue? the response alternative or don't you have an opinion on this issue is an explicit “don't know” or “no opinion” option. When such options are provided as part of the question, the percentage of respondents who choose this alternative is higher than when it is not present, and respondents have to volunteer that they “don't know.”

...