The rescission of behaviors: Inconsistent responses in adolescent sexuality data

摘要: In longitudinal survey research, certain questions can be rescinded illogically. For instance, respondents who at Time 1 report having had sexual intercourse may 2 never done so. This paper reports measurement techniques and analyses of these types inconsistencies from an ongoing adolescent sexuality project. Inconsistencies in intercourse, masturbation, other behaviors are reported compared to rates studies less sensitive within the same study. Three conclusions presented: (1) should considered a natural part any process incorporated into response model; (2) inconsistency particular data support contention that appropriate quality for analytical purposes obtained; (3) fact contain substantive information concerning processes under consideration.