THE ROLE OF SCHOOL CONTEXT ON AGGRESSION AND SUBSTANCE USE DURING MIDDLE SCHOOL: A PERSON-ORIENTED DEVELOPMENTAL APPROACH

Despite more than 30 years of research and a growing list of evidence-based prevention strategies, high rates of risk behavior such as aggression and substance use continue to persist among early adolescent urban boys (Hawkins et al., 2015). Much of the early research and intervention and prevention strategies have focused on individual and then family factors related to risk, with less attention paid to other contextual factors, particularly school (Bosworth, 2015; Petras & Slobada, 2014). Practice strategies that integrate knowledge regarding developmental trajectories, contextual malleable risk, and protective factors are needed to better address problem behaviors. Using data from a sample of urban middle school boys in four US cities, risk patterns based on social-cognitive deficiencies (i.e., learning, social skills, anxiety and conduct problems) at school entry and their relation to aggression and substance use trajectories through the middle school years are examined. In addition, the role of school context and the interaction between patterns of social-cognitive risk and developmental patterns of aggression and substance use is investigated. Informed by person-oriented developmental psychopathology principles, the aim of this dissertation is to evaluate the role of the school environment in influencing youth aggression and substance use across profiles of social-cognitive risk. The study shifts the focus from risk and intervention based in the individual to factors that influence risk and development salient in the school context. Results from this study can inform school-based behavioral prevention and intervention programming. Using latent class analysis, four patterns of youth risk profiles were identified in the sample: a “low all” group (61.3%) which displayed low risk across all factors, a group with “moderate” learning and anxiety problems (15.5%), a group with “poor social skills” that showed some learning and conduct problems (16.9%), and a “high all” group with high-levels of risk across all characteristics assessed (6.3%). Latent growth curve analyses were conducted and revealed that the “low all” group had the lowest growth of aggression and substance use through middle school. The “moderate” group had an average growth of behaviors, while the “poor social skills” group had higher levels of aggression and substance use at 6th and 7th grade. The “high all” group showed the fastest growth in problem behaviors and had the highest levels of aggression and substance use at 8th grade. Analyses conducted to evaluate the role of school-level characteristics (i.e., quality of student-student and student-teacher relationships, sense of safety, awareness of school problems and school norms about behavior) identified two significant school effects on behaviors. Quality of student-student relationships significantly moderated the relation between the probability of membership in the “poor social skills” group and growth of relational aggression only. Additionally, quality of student-teacher relationships moderated the relationship between membership in the “high all” group and growth of non-physical aggression only. Further analysis found that higher quality student-student relationships among the “poor social skills” group were associated with a decrease in relational aggression by an effect size of d = 0.24 (p ≤ 0.001). Moreover, higher quality student-teacher relationships were associated with a reduction in non-physical aggression among the “high all” group by an effect size of d = 0.25 (p ≤ 0.05). Notably, no school factors moderated substance use across all groups. There are two important implications of these findings. First, social-cognitive characteristics of boys at school entry are related to growth in risk behaviors over the course of middle school. These findings point to important variation in risk during this developmental period. The findings also highlight the role of school context for some behavioral risk outcomes, particularly for non-physical and relational aggression. Results are relevant to prevailing national prevention strategies and educational trends that focus on a multi-tiered approach to behavioral risk management. The dissertation concludes with a discussion of how future research can promote the development of practice strategies and interventions that target pattern-specific behavioral risk among vulnerable urban youth populations