Recanting of substance use reports in a longitudinal prevention study

We analyzed recanting of substance use reports for lifetime use of alcohol, alcohol to get drunk, cigarettes, marijuana and cocaine in an 8-wave panel study designed to evaluate the Drug Abuse Resistance Education program in the State of Illinois. Although this phenomenon has been identified elsewhere, the current analysis of recanting is a unique attempt to track this behavior over the entire course of adolescence. Overall, rates of recanting for specific drugs were extremely high, ranging from 45% for lifetime reports of alcohol use to 81% for lifetime reports of cocaine. Most recanting occurred in the wave immediately following the wave of first disclosure. Paralleling results from other studies, race/ethnicity was an important correlate of recanting in both bivariate and multivariate analyses. African American respondents had higher rates of recanting than White subjects. Even after controlling for the number of follow-up waves, the later the wave of first disclosed lifetime drug use, the lower the probability that drug use would be recanted ever (for all substances) or in the wave immediately following first disclosure (for reports of ever having been drunk or for lifetime marijuana or cocaine use). Alternative causes for this phenomenon are discussed. Implications for the design and interpretation of multiwave school-based panel surveys targeted toward adolescents are also addressed. (C) 2003 Elsevier Science Ireland Ltd. All rights reserved.