Differences in the Validity of Self-Reported Drug Use Across Five Factors: Gender, Race, Age, Type of Drug, and Offense Seriousness

Abstract

This study expands our knowledge about the validity of self-reported drug use by examining how gender, race, age, type of drug, and offense seriousness interact to affect the validity of self-reported drug use. This study also provides a conceptual framework that can be used to examine the validity of self-reported drug use. Differences in the validity of self-reported drug use are explained by examining differences in underreporting and overreporting. Differences in underreporting and overreporting are then further examined while controlling for differences in base rates of drug use. As shown, whether one controls for base rates of use may drastically affect estimates of underreporting and overreporting. By using hierarchical loglinear, logit, and logistic regression models with the Drug Use Forecasting data, we show that Black offenders provide less accurate self-reports than White offenders. Black offenders do so because they are more likely to underreport crack/cocaine use than White offenders. This difference, however, disappears once differences in base rates are controlled. A Black offender who tests positive is not more likely to underreport crack/cocaine use than a White offender who tests positive. Black offenders are also more likely to overreport both marijuana and crack/cocaine use relative to White offenders. Contrary to the first, this difference is not attributable to a difference in base rates. Methodological and substantive implications of this distinction are discussed. No differences across gender, age, or offense seriousness were found.