categorical and continuous variables, respectively. Although tests of sociodemographic characteristics also included the comparison of CUD only vs. ND only groups, results from this comparison are provided for descriptive purposes only. Substance use characteristics, psychiatric disorders, and psychosocial and physical health correlates were outcome variables in logistic and linear regression models, with the 3-group CUD/ND variable (i.e., current CUD + ND vs. CUD only vs. ND only) as the explanatory variable. Models were adjusted for age, gender, race, and education. Logistic and linear regression analyses yielded adjusted odds ratios (AORs) and adjusted unstandardized partial regression coefficients (b weights), respectively. Standard errors for all estimates were adjusted for NESARC design characteristics using SUDAAN (SUrvey DAta ANalysis).
