Despite the ASI's wide use in substance abuse treatment and research, factor analysis on the ASI is sparse and controversial. Innovative statistics and programming strategies will be applied to and explored on categorical response formats and causal indicators, both either inappropriately or rarely considered in previous research. ASI's measurement invariance will be evaluated to examine whether any measurement bias exists across different groups, such as ethnicity/race, age, gender, primary drug of choice for participants and geographic region from which data were collected. The central hypothesis of this proposal is that a stable latent structure will emerge after the measure is subjected to state of the art statistical analyses that are suggested for establishing measurement equivalence. Some parameters (i.e., loading, intercept, threshold, error variance, factor mean, and factor variance/covariance) of this structure are expected to differ across participants from diverse backgrounds. That is to say, various populations' ASI values can be compared on a common scale based on a stable latent structure with parameters of values specific to each population. The proposed study is the first attempt, to our knowledge, to collectively evaluate ASI- Lite's levels of invariance across various subpopulations at baseline and prepare for longitudinal invariance analysis at multiple time points. The proposed application is also innovative because it proposes to use the latest available analytic strategies applicable to the ASI. Finally, this study proposes to use the largest available ASI dataset from the Treatment Research Institute's national Drug Evaluation Network Study and recently collected data from the NIDA Clinical Trial Network. The CTN consists of multiple community-based research protocols, most of which use the ASI-Lite. Moreover, approximately half of the first 20 protocols have completed and locked datasets that are readily accessible for analysis. The results of the proposed study are of urgent importance because they provide an empirical basis for the upcoming comparisons across treatments that are included in each protocol. The interdisciplinary research team includes statisticians, an ASI expert, a substance abuse researcher, and community reviewers. 1 Categorical items, causal indicators, and levels of invariance have not been adequately considered in previous latent structure studies on the Addition Severity Index. The proposed study is the first attempt, to our knowledge, to collectively build the ASI-Lite's latent structure to encompass indicators in mixed formats and to evaluate its levels of invariance across various subpopulations at baseline. ? ? ?
