Most variance in use and misuse of various substances is accounted for by common (rather than substance-specific) genetic and environmental factors (Hall & Lynskey, 2005; Kendler et al., 2003). Nevertheless, much of the literature treats substances as distinct entities (e.g. Kahler et al., 2004; Kashdan et al., 2005; Saha et al., 2006; Wallace et al., 2003). We evaluated the utility of modeling involvement with multiple substances in terms of a unifying poly-substance dimension. Specifically, we evaluated a dimensional conceptualization of poly-substance use in a longitudinal study that followed male and female participants from ages 11 to 17. To understand the construct validity of this dimensional conceptualization, we modeled the impact of age and sex on the properties of the dimension. Rather than impacting specific substances individually, age and sex may impact properties of a unifying substance involvement dimension (e.g., being male may increase risk for involvement with multiple substances, which could be modeled as a sex effect on the entire dimension, transcending specific substances). This would support conceptualizing adolescent substance use in terms of a unifying poly-substance dimension, and focusing
