3-058 - How Do School-Based Interventions Improve Children’s Academic Development? Estimating Mediation Effects with Causal Methods

Session Type: Paper Symposium

Integrative Statement

School-based interventions in K – 12 settings, increasingly targeted at low-income, racial/ethnic minority students, use a variety of strategies to promote children’s academic and social-emotional development. Implementation of such programs is typically guided by complex, multi-level developmental theories, which identify classroom and child-level mediators explaining program impacts on setting- and individual-level outcomes. Yet, few studies have explicitly examined such mediating processes using rigorous modeling approaches that maintain the internal validity of randomized study designs. The proposed symposium will fill this gap by using causal methods to test hypothesized pathways linking intervention to positive classroom and child-level outcomes.

Study 1 examines classroom-level program impacts with an instrumental variables approach, finding that the effect of RULER on classroom organization was mediated by improvements in emotional support. Study 2 uses instrumental variables and inverse probability weighting to find that impacts of the INSIGHTS intervention on first grade math skills were mediated through improvements in classroom emotional support and organization. The third study also uses instrumental variables and finds that impacts of the 4Rs program on children’s math achievement were mediated through reductions in children’s hostile attributional bias. Finally, Study 4 examines adolescents in the Every Child, Every Day program and did not find rigorous evidence in support of the hypothesis that impacts on math achievement would be mediated by improvements in instructional rigor.

Findings suggest some rigorous evidence for developmental theories linking school-based interventions to children’s academic development. Results will be discussed with implications for developmental theory, intervention design, implementation, and scale-up.

Sub Unit

• Education, Schooling

Chair

• Stephanie M Jones, Harvard Graduate School of Education

Individual Presentations

• Using Instrumental Variable Techniques to Causally Test the Theory of Change of a School-based Intervention - Presenting Author: Catalina E Torrente, Yale University; Christina Cipriano Crowe, Yale University; Lori Nathanson, Yale University; Susan E Rivers, Yale University; Marc A Brackett, Yale

• Testing Classroom Mechanisms Linking Social-Emotional Learning and Academic Development: Evidence from a Randomized Trial - Presenting Author: Meghan Patricia McCormick, New York University; Elise Cappella, New York University; Erin O'Connor, New York University; Sandra G McClowry, New York University

• Individual and Classroom Mediators of the 4Rs Program: Examining Mechanisms with Instrumental Variables Estimation - Presenting Author: Andres Molano, Universidad de los Andes, Bogota, Colombia.; Stephanie M Jones, Harvard Graduate School of Education; Joshua L Brown, Fordham University; J. Lawrence Aber, New York University

• Testing Classroom Instructional Quality as a Causal Mediator of an Instructional Improvement Intervention on Math Achievement - Presenting Author: Juliette Berg, University of Virginia; J. Lawrence Aber, New York University; Diane Early, University of Rochester; Stacey Alicea, New York University; Edward Deci, University of Rochester