This five-year investigation is designed to provide a more refined understanding of the risk factors and risk mechanisms involved in social deviance. Both etiology and consequence models are posited including specific tests of direct, moderator, and mediating effects. A unique feature of this study is a reliance on data from two generations of participants including a G1 cohort followed annually in the 7th through 9th grades (a 50 percent sample of the 7th graders in the Houston Independent School District in 1971) again as young adults in their 20s, and most recently when they were 35-39 years of age; and a cohort of N=2,477 youth who are the children (G2) of the G1 cohort and who were interviewed between 1994 and 1998 when they were 12 or 13 and will be re-interviewed between 2003-07 when they will be 21 or 22. A rich vein of psychosocial (e.g., violations of normative role expectations, self-rejection, deviant motivations, internal and external controls, and peer and parental rejection, socialization, and modeling), behavioral (e.g., drug use, violence, and criminal activities), psychological (e.g., psychiatric disorders), and other relevant theoretical constructs tied to social control, differential association, social learning/influence and labeling formulations are modeled. Additional structural variables reflecting residential (e.g., neighborhood disorganization) and school contextual influences are collected from the US Censuses of 1980 (before G2 subjects were born), 1990 (when G2 subjects were 4-9) and 2000 (when G2 subjects were 14-19), and published aggregate data for schools attended by G2 subjects. A series of integrated multivariate analyses are used to examine inter- and intragenerational influences including establishing construct factorial (i.e., measurement) and structural regression coefficient equivalence across generations. Models are estimated using multi-group structural equation modeling with latent variables, multivariate ordinary least squares and logistic regression with interaction terms, and multilevel modeling to control for potential clustering influences originating from intact social groups (i.e., families or schools). The model of social deviance elaborated and tested has yielded to date promising findings regarding the role of affect, cognition, personality, and social influences in the generation of delinquency and drug use.
Showing the most recent 10 out of 15 publications
