Inaccurate self-reporting is the primary threat to the validity and utility of self-report in the research on illicit drug use. The purpose of this study was to estimate the sensitivity and specificity of self-report of drug use, using urine toxicity screening as a criterion variable, and to explore the individual characteristics associated with false reporting in a sample of urban African American males. Baseline data from urine tests and a questionnaire about self-report of illicit drug use were obtained from 290 hypertensive African American males enrolled in an ongoing hypertension clinical trial. Sensitivity and specificity of self-report were estimated, and multivariate logistic regression analysis was conducted to examine factors associated with true or false reporting of drug use.
Result: The sensitivity and specificity of self-report were 0.87+/-0.019 and 0.63+/-0.02, respectively. Members of the underreporting group were more likely to be older, more educated, employed, living with family, and having experience with jail. Jail experience (OR, 2.0) and living with friends or family (OR, 1.7) were the two strongest predictors of false reporting. This study further confirms that social desirability and fear of the consequences of drug use are major contributing factors in underreporting of drug use.