(no) and 1 (yes). Scores were summed to create variables for maternal involvement and paternal involvement. The parent-child communication scale contained 3 items asking adolescents whether or not they and their father/mother figure talked about news, their problems, or other things like movies, friends or anything else. Response options were 0 (no) and 1 (yes) for each item. Sum scores were used to create variables for mother-child communication and father-child communication. The parent-child closeness scale included 2 questions that asked participants how well they got along with their father/mother figure most of the time and how close they felt to their father/mother figure. Response options ranged from 1 (poor) to 4 (excellent), and from 1 (not at all close) to 3 (very close) for the two questions, respectively. Scores were standardized and averaged across the two items to create variables representing father-child closeness and mother-child closeness. Correlations between the three scales ranged from .39 to .49 (mean r = .42) and from .50 to .57 (mean r = .54) for fathers and mothers, respectively. Some adolescents reported parenting behaviors for their non-biological father (n = 370) or mother (n = 82) figures. Because we only had data on alcohol dependence
