Category: Addictive Behaviors

PS15- #A6 - Making Decisions With Trees: Identifying Predictors of Marijuana Outcomes Among College Students Using Recursive Partitioning

Sunday, Nov 19
9:00 AM – 10:00 AM
Location: Indigo Ballroom CDGH, Level 2, Indigo Level

Keywords: Addictive Behaviors

Exploratory analyses are needed to identify unique constellations of variables that are predictive of marijuana use outcomes among college students. The current study examined over 100 predictors/antecedents (e.g., impulsivity, identity, motives, norms, and protective behavioral strategies) of college student lifetime marijuana use, past month marijuana use, and negative marijuana-related consequences using recursive partitioning. Recursive partitioning, a machine learning approach to exploratory analysis, is particularly well suited to developing parsimonious predictive models from behavioral data (Hellemann et al., 2009). Though not without drawbacks (King & Resick, 2014; Marshall, 2001), tree-based modeling has advantages over traditional correlational and regression-based operations; namely, tree-based modeling does not rely on assumptive statistical tests and effects determined by p-values less than .05. Additionally, tree-based modeling is capable of parsing the predictive utility of large numbers of variables in big datasets, such as the one collected by the Marijuana Outcomes Study Team in the present study. Participants (N = 8141) completed measures online across 11 sites in the United States. Results from the recursive partitioning analyses identified unique predictors of marijuana outcomes including injunctive norms, internalization of college marijuana use culture, marijuana identity, descriptive norms, use of protective behavioral strategies, and positive urgency. Our results lend empirical support to intervention approaches targeting these predictors of marijuana outcomes (especially use of protective behavioral strategies), and demonstrate the promise of tree-based modeling to assist with clinical decision-making.

Adam Wilson

Doctoral Student
University of New Mexico

Kevin Montes

Center on Alcoholism, Substance Abuse, & Addictions, University of New Mexico

Adrian Bravo

Center on Alcoholism, Substance Abuse, & Addictions, University of New Mexico

Bradley T. Conner

Associate Professor
Colorado State University
Fort Collins, Colorado

Matthew R. Pearson

Research Assistant Professor
University of New Mexico