Presenter:
Brian R. Flay, Oregon State University
Carol G. Allred, Positive Action, Inc.
David DuBois, University of Illinois at Chicago
Purpose: Many students lack the necessary social skills and character needed to succeed academically and to avoid involvement in anti-social behavior. The project is evaluating a comprehensive program, Positive Action (PA) that previous quasi-experimental evaluations suggest produces substantial gains in prosocial behavior and school performance. The primary purpose of the project is to conduct a randomized efficacy trial of PA.
Population: Schools were selected from clusters of similar Chicago Public Schools (stratified on ethnic distribution, % free/reduced lunch, mobility, attendance and truancy rates, standardized test scores). Study schools consist of an average of 2-3 classrooms per grade with an average mobility rate of 12.5% (so that 500 students will stay for the three years of the study).
Methods: Fourteen schools were assigned randomly from matched pairs to receive the intervention or traditional instruction. Cross-site data collection includes measures of social competence, character, behavior and achievement. Site-specific data includes extensive process measures of fidelity of implementation and dosage of exposure for all components of the PA program as well as additional measures of hypothesized mediators of the intervention's effects. We will use hierarchical statistical models (random regression models and GEE) that can accommodate longitudinal data with nested observations and missing observations.
Intervention: Positive Action includes five major components: 1) Classroom Curricula for every grade; 2) Principal's Kit (School Climate); 3) Family Kit; 4) Teacher/Staff Training; and 5) Community Involvement Program. It is a comprehensive program designed to change school context, increase parental involvement, and improve students' self-concepts.