As described above, a supraordinate substance use lifestyle factor, built from the eight indicators representing adolescent marijuana use and friends’ support of substance use, was used due to high level of multicollinearity between adolescent marijuana use and friends’ support of substance use (r = 0.84). This alternative model (Fig. 3) had better BIC (17221.81) than did the first model (BIC = 17225.23). Except for a significant χ2 value, fit indices for this alternative model were adequate: χ2(81) = 316.38, p < 0.001, CFI = 0.93, RMSEA = 0.05 (90% C.I. = 0.05 to 0.06), SRMR = 0.05. In this alternative model, all indicators were significantly related to the latent variable to which they were assigned at p<0.001. Substance use lifestyle in adolescence was a strong, significant predictor of problematic marijuana use in early adulthood (p<0.001), and effortful control was a significant, negative predictor (p<0.01), although its effect size was in the small range (Cohen 1988). An interaction model for marijuana use was estimated by adding an interaction term between effortful control and the substance use lifestyle factor. The interaction term was not a significant predictor of early adulthood problematic marijuana use.
