Advancing the Science and Practice of Social and Emotional Learning (SEL) in Schools: A Closer Look at the Evidence, Measurement Approaches, and the Precision of Classroom Observation Tools

Abstract

Over the past 20 years, social and emotional learning (SEL) has become increasingly present in children’s daily school experiences, particularly in response to reported declines in children’s mental health and well-being due to the COVID-19 pandemic. Despite the well-identified need for school-based social and emotional skill-building approaches, critics of SEL question its place in schools, the rigor of existing empirical work, and current measurement approaches. With the growing demand for SEL in schools despite many critics, it is an ideal time to take stock of the evidence and measurement approaches that underlie research and practice in SEL. The first two papers in this dissertation employ data from a recently published meta-analysis of SEL intervention impacts, which documented improvements on a host of outcomes, including a general category of SEL skills (Cipriano et al., 2023). In the first paper, I conduct a meta-analysis examining the impacts of SEL interventions on specific SEL skill domains, organized by the two leading frameworks in SEL research and practice. In a sample of 59 studies from the United States involving 25,177 K-12 students, I find that children who participated in school-based SEL interventions showed statistically significant improvements in multiple SEL skill domains relative to their peers in the control group. In the second paper, I code 238 measures of SEL skills from the meta-analysis by SEL skill domain, measure type, developmental stage, and presence of internal consistency statistics. In the third paper, I apply Generalizability Theory (G Theory) to a classroom observation tool, the Adapted Teaching Style Rating Scale (A-TSRS), designed to capture the classroom structures and teaching practices that support children’s social and emotional development. Specifically, I use G Theory to decompose the sources of variance contributing to imprecision in scores on the A-TSRS and present reliability coefficients for a variety of potential observation scenarios. I find that substantial gains in precision arise from adding multiple observations within the same day. Taken together, the three papers in my dissertation provide a clear and transparent perspective on the opportunities and challenges facing the field and present ideas to advance the science and practice of social and emotional learning.