Abstract
AIM: This study applied both item response theory (IRT) and multiple indicators-multiple causes (MIMIC) methods to evaluate item-level psychometric properties of diagnostic questions for hallucinogen use disorders (HUDs), differential item functioning (DIF), and predictors of latent HUD. METHODS: Data were drawn from 2004-2006 National Surveys on Drug Use and Health. Analyses were based on 1548 past-year hallucinogen users aged 12-17 years. Substance use and symptoms were assessed by audio computer-assisted self-interviewing methods. RESULTS: Abuse and dependence criteria empirically were arrayed along a single continuum of severity. All abuse criteria indicated middle-to-high severity on the IRT-defined HUD continuum, while dependence criteria captured a wider range from the lowest (tolerance and time spent) to the highest (taking larger amounts and inability to cut down) severity levels. There was indication of DIF by hallucinogen users' age, gender, race/ethnicity, and ecstasy use status. Adjusting for DIF, ecstasy users (vs. non-ecstasy hallucinogen users), females (vs. males), and whites (vs. Hispanics) exhibited increased odds of HUD. CONCLUSIONS: Symptoms of hallucinogen abuse and dependence empirically do not reflect two discrete conditions in adolescents. Trends and problems related to hallucinogen use among girls and whites should be examined further to inform the designs of effective gender-appropriate and culturally sensitive prevention programs.