EFFECTS OF INTERVIEWER GENDER, INTERVIEWER CHOICE, AND ITEM WORDING ON RESPONSES TO QUESTIONS CONCERNING SEXUAL-BEHAVIOR

We examined factors influencing responses to questions on sexual behav ior among adult respondents 18-49 years old (unweighted N = 2,030) obt ained through a random-digit dialing survey. Based on self-disclosure and perceived control theory, we hypothesized that giving people a cho ice in selecting the gender of their interviewer rather than being ass igned an interviewer, and using questions that are ''supportive'' of w hat may be perceived of as nonnormative behavior (enhanced items), wou ld increase data quality relative to, respectively, matched- or opposi te-gender interviewer conditions and standard worded items, The enhanc ed items facilitated responding to a number of sensitive topics, Howev er, the effects of item wording on item response are often mediated by interviewer conditions. The ''choice'' results suggest that giving re spondents greater control decreases question threat. However, the over all findings argue for matching respondents and interviewers on gender over opposite-gender interviewers or allowing respondents to select t heir interviewer's gender. Wording and interviewer manipulations reduc ed the discrepancies between men's and women's self-reports of sexual behavior, but they did not eliminate them, and in some cases they had no effect. The present findings suggest that males tend to be influenc ed by variations in item wording, interviewer gender, and respondent c ontrol across a somewhat wider range of sexual topics. In general, the findings recommend matching respondents and interviewers on gender an d the use of more supportive wording in sexual behavior questions. How ever, for assessment of some topics (e.g., sexual violence) in particu lar segments of the population (e.g., men), other procedures, such as increasing respondent control, may be a better choice. Overall, the da ta support the view that in terms of preferred procedures, not all sex ual topics are created equal.