215-221 Lauritsen

may be many years in the past. The accuracy of reporting on sensitive matters relating to sexual behavior can also shift as the culture changes, and that accuracy may vary with age. Because, by definition, self-reports of the initiation of sexual activity cannot be externally validated, the scientific literature has necessarily focused on the consistency and reliability of reporting, through two broad analytic strategies. The first examines the consistency within data reported by the same individual at different times.2 The relatively few studies on this topic have found that some participants provide contradictory information over time about their own sexual experiences, and that these inconsistencies are correlated with race and gender. These findings suggest that errors in the measurement of sexual activity are not random and, consequently, that differences in sexual behavior by subgroups may be consistently biased across data sources. The correlates and consequences of inconsistency in reporting warrant more extensive investigation for this reason. The second strategy for testing the reliability of self-reporting examines aggregate consistency across multiple national data sources. These comparisons have suggested that American women have reliably reported their age of sexual initiation.3 However, this strategy has several limitations. In most cases, the survey data were obtained from retrospective questions Janet L. Lauritsen is associate professor in the Department of Criminology and Criminal Justice, University of Missouri-St. Louis, and C. Gray Swicegood is associate professor at the Department of Sociology, University of Illinois-Urbana, Ill. The authors thank Gillian Stevens, Seymour Sudman and Constance Nathanson for helpful comments on earlier versions of this article. The National Youth Survey data were obtained through the Inter-University Consortium for Political and Social Research at Ann Arbor, Michigan. The Consistency of Self-Reported Initiation of Sexual Activity