Knowledge Question

A knowledge question is designed to capture the extent to which people have stored factual information in long-term memory and how well they can retrieve and respond with that information when asked a survey question about a given topic. Knowledge, as a concept, is distinct from attitudes and opinions. Knowledge, moreover, is not synonymous with the term information. Whereas information describes a wider breadth of content that may include non-neutral as well as neutral elements, knowledge is considered neutral, factual content. The term factual is one that has been debated. Facts refer to those content elements on which there is consensual agreement. For example, it is a fact that it takes two thirds of Congress to override a presidential veto. Whether or not a given politician was arrogant during a debate, however, is a subject that is not likely to yield consensual agreement. Many topics can be assessed with knowledge questions, including politics, health, consumer issues, popular culture, science, and education.

Knowledge questions are used for several purposes. First, knowledge questions can be used to screen people for follow-up questions about people's attitudes on a given topic. If it is determined that a person has little or no knowledge of a topic, it may not be efficient for the researcher to ask that person follow-up attitude questions on that subject. Second, knowledge questions are used to assess a person's intellectual engagement on a topic, because knowledge is often predictive of various attitudes and behaviors. For example, citizens who are knowledgeable about politics tend to be the citizens who vote in elections; in other words, political knowledge predicts voting. Third, knowledge questions may be used for evaluative purposes. A school's performance may be assessed by how well its students have mastered a given set of material.

There are two primary types of knowledge questions: recognition and recall. Awareness questions are often assessed with recognition measures. Recognition measures ask people whether or not a given person, topic, or event is familiar to them. Specific knowledge content, however, is often assessed through recall measures. Recall items come in two forms: aided and unaided. An aided recall question is a closed-ended question that presents a respondent with several response choices from which the respondent selects an answer. An unaided recall question is asked in an open-ended format; the respondent receives no hints about the answer, and the respondent's answer is often recorded verbatim. Assessing a respondent's general awareness of a person, topic, or event is often less cognitively taxing on him or her than assessing a respondent's memory of specific knowledge content.

The main concern over recognition items is social desirability bias. People may say that they recognize an object because they want to appear informed. Consequently, some researchers follow up recognition items with specific content questions or include false [Page 412](bogus) names or events in their recognition items to determine the extent to which people provide false positives (errors of commission) when it comes to recognition.

A general concern over asking knowledge questions is that respondents may feel intimidated when they perceive that they are being tested; they may be afraid of giving the incorrect response and looking ignorant or foolish. Anxiety can sometimes be minimized by prefacing questions with phrases such as Do you happen to know … or As far as you know…. Prefacing knowledge questions with softening phrases, however, may make people feel too comfortable giving "don't know" responses when they do know the correct answers but are not confident they know them and thus are hesitant to take a chance answering the questions. This can be problematic for researchers who are assessing the knowledge levels among different groups, particularly if some groups have greater propensities to guess than other groups do. For example, research has long shown that women are less likely to guess at knowledge items than are males. It is difficult to sort out a group's propensity to guess from its true levels of knowledge acquisition.

...