Empowering Randomized Intervention Studies on Students’ Socio-Emotional Learning Outcomes: A Database of Design Parameters for Power Analysis

Introduction. Determining which interventions meaningfully foster students’ socio-emotional learning (SEL) remains an urgent concern in educational and psychological research. Valid statistical conclusions about an intervention’s effect, however, require robust experimental designs with adequate sample sizes. Therefore, the availability of appropriate and reliable estimates of design parameters for a priori power analyses is crucial.Methods. To generate a comprehensive database of design parameters for SEL research, we analyzed data from nine German probability samples involving students in grades 1 to 13. These data covered self-oriented, other-oriented, and task-oriented domains of SEL outcomes, assessed via student self-report, parent rating, and teacher rating methods. Using linear and multilevel regression modeling techniques, we estimated the following design parameters: Intraclass correlations at the classroom and school level as well as the proportions of variance explained by different sets of covariates (sociodemographic characteristics, baseline measures, their combination) at the student, classroom, and school level. We also derived meta-analytic summaries of these design parameters, broadening their applicability.Results. The compiled SEL design parameters along with meta-analytic summaries span both the overall student population and subpopulations defined by academic tracking. This provides a flexible foundation for a range of study designs, including single-level (non-clustered), two-level (students nested within schools), and three-level (students nested within classrooms nested within schools) randomized experiments.Discussion. The present database, complemented by extensive supplemental material, serves as a curated and practical resource for planning sufficiently powered randomized intervention studies targeting students’ SEL outcomes.