Opinion Question

Opinions that individuals hold about an issue are potentially quite complex, considering how opinions comprise beliefs, feelings, and values on that issue. As a result, survey questions designed to assess a respondent's opinion on an issue can tap a combination of the respondent's feelings and thoughts. As composite measures, however, opinions are gauged through a variety of opinion items.

The most basic opinion question, sometimes called an “attitude question,” is designed to measure the direction of opinion. That is, where does the respondent stand on the issue or attitude object? Such opinion items, typically closed-ended, can be dichotomous in nature (e.g. Do you support or oppose abortion? Answer categories: Support; Oppose. Are you in favor of the death penalty for a person convicted of murder? Answer categories: Yes; No). Opinion items [Page 551]also can employ a scale that measures not only the direction of one's opinion but also the respondent's intensity of opinion (e.g. To what extent do you support or oppose abortion? Answer categories: Strongly oppose; Somewhat oppose; Somewhat support; Strongly support).

Open-ended questions can be used as well to measure opinions, but these can be more effective in shedding light on why respondents hold the opinions they hold. For instance, respondents in the American National Election Studies who express their intention to vote for a particular candidate are asked, Is there anything in particular about [Politician X] that might make you want to vote for him? Other open-ended items, such as the “Most Important Problem” question (What do you believe is the most important problem facing the nation today?), do not offer respondents an opportunity to articulate reasons why they volunteered a particular issue. However, like the aforementioned opinion questions, responses to Most Important Problem items can be used to shape public policy, campaign efforts, or marketing strategies.

Regardless of the issue under study, how an opinion question is asked can have a significant impact on its response. Hence survey researchers routinely take into consideration a number of concerns when crafting their questionnaires. First, are nonattitudes a potential problem? Also known as “false positives” (i.e. errors of commission), nonattitudes occur when respondents report an opinion on an issue about which they know nothing or really have no attitude. To circumvent the problem of nonattitudes, survey researchers can employ a filter question immediately before the opinion question such that only respondents who report knowledge or awareness of the issue are asked their opinion about that issue. Another way to reduce nonattitudes is to offer respondents a “Don't Know” option or a middle response category (e.g. “Neither support nor oppose”). Though common, the inclusion of a middle response category can have the unintended consequence of generating “false negatives” (i.e. errors of omission), the reporting of no opinion when in fact one exists.

A second consideration in the crafting of opinion items concerns how the question and response categories are worded. The question should include a clearly specified attitude object and should not be double-barreled; put another way, respondents should not be asked to express their opinion about two attitude objects (e.g. Do you support abortion and health care reform?). In addition, questions should not include double negatives, colloquialisms, or leading terms such that respondents feel pressured to provide a socially desirable answer. Response alternatives, the construction of which should follow these same guidelines, also should be balanced and include a sufficient range of variation.

...