We have developed a dynamic developmental systems (DDS) framework (Capaldi et al., 2005) that seeks to explain onset and changes in behaviors such as substance use from (a) general developmental risk pathways (e.g., parental monitoring and poor discipline, boy’s antisocial behavior and deviant peer associations, boy’s depressed mood) and (b) social influences from key developmental interactants that are specifically related to the outcome under study (e.g., use of alcohol by parents and peers). The goal of the current study was to examine prediction models of growth in alcohol use for the middle and high school periods, spanning ages 11–12 to 17–18 years, using the Oregon Youth Study (OYS) sample of boys, who were at risk for conduct problems because of their disadvantaged neighborhoods. Models were tested by using a two-part random-effects model for semicontinuous data (Olsen and Schafer, 2001), which has been used in recent studies of alcohol use at adolescence (e.g., Blozis et al, 2007; Brown et al., 2005). Part I involved prediction to use vs. nonuse; Part II involved prediction to yearly volume given use.
