ABSTRACT
The frequency with which a patient experiences physical symptoms is of considerable interest in medical research and practice. Moreover, many studies include questions about the frequency of health relevant behaviors, ranging from the number of cigarettes smoked to the weekly hours of exercise or the number of sexual partners. Although these questions may seem straightforward at first glance, answering them poses considerable cognitive demands. In this chapter, I review the cognitive tasks involved in answering behavioral frequency questions and highlight how apparently formal features of the questionnaire may determine respondents’ reports. In clinical practice, health care providers are likely to draw on additional information about the patient, thus limiting the practical relevance of questionnaireinduced context effects. In contrast, self-reports in response to standardized ques-tions are the primary source of information in many epidemiological studies, including studies that compare the health status of younger and older adults. To the extent that features of the questionnaire have a differential impact on older and younger respondents, they may lead to inappropriate conclusions about age-related differences in the behaviors studied.
