Bivariate associations between latent class of psychopathology and demographic, early environment, trauma, perceived substance use by sibling and friends, parent substance use, and other psychiatric disorder variables were examined using chi-square tests, controlling for the clustering of family data. Variables with significant bivariate associations (p ≤ .05) were included in a series of 3 multinomial logistic regressions, all adjusted for age, gender and race and constructed according to familial risk, family environmental factors and perceived substance use of friends. Model 1 tested the association between the latent classes and the following familial risk factors: 4-level familial risk variable, maternal marijuana abuse/dependence, paternal smoking, paternal externalizing disorder, maternal depression and sibling alcohol and drug use. Model 2 added the following early environmental factors: father marital status, mother rule consistency, closeness to father and mother and childhood sexual or physical abuse. Model 3 added perceived substance use of friends. Examination of the tetrachoric correlations between covariates indicated no evidence of collinearity in the regression models. Dummy variables representing missing maternal data (for mothers who did not complete an interview) and missing sibling
