Key points are not available for this paper at this time.
IN THE past 30 years, it has become commonplace to collect interview and questionnaire data about the personal sexual behavior of respondents, and most researchers have come to take it for granted that if the questions are asked in a reasonable and non-threatening way, they will be answered. The problem which continues to nag is the matter of validity of the response. Did the respondent actually do what he said he did? In methodological discussions, it is usually recommended that questions be non-threatening and that rapport be established,1 but the contribution of rapport to response validity is not known. Kinsey et al. recommend placing the burden of denial of an act on the respondent by, for example, asking when he first engaged in an activity, not whether.2 Rapid-fire questioning with built-in cross-checks are also recommended for the interview situation,3 but regular schedule. Thus there has appeared to be no way of determining the accuracy of reported sexual behavior.
Udry et al. (Tue,) studied this question.