Machine Learning-Assisted Evaluation of a School Social Work Intervention for Students at Risk of Dropping Out: Effects on School Belonging and Academic Engagement

Abstract

The purpose of this study was to assess the efficacy of a school social work intervention for students at risk for dropping out of school with a special focus on its impact on students' school belonging and academic engagement. The study sample comprised of 130 students who were in Grade seven and were considered high-risk of school dropout, with 65 students in the experimental group that received school social work intervention and 65 students in the control group that received regular school services. The design of quasi-experimental pre-test/post-test control group was used for quantitative approach. Two instruments—the 15-item School Belonging Scale (Peer Relations, Teacher Relations, and Emotionally Connected to School) and the 18-item Academic Engagement Scale (Behavioral Engagement, Emotional Engagement, and Cognitive Engagement)—were used to collect data. Means and standard deviations, analysis of covariance (ANCOVA), and partial eta squared were used to analyze the data. The results indicated that before the intervention the students had moderate scores for school belonging (M = 2.48, SD = 0.65) and academic engagement (M = 2.39, SD = 0.61). The experimental group's mean increased from 2.48 to 3.82 on the school belonging measure after intervention, while the mean change for the control group was 2.45 to 2.50. For the ANCOVA results, there was a significant group effect (F (1, 57) = 96.85, p < .001, partial η² = .63). In the same way, the academic involvement score rose from 2.39 to 3.79 for the experimental group while that of the control group had a marginal improvement of .1, from 2.37 to 2.42. The difference was statistically significant, F (1, 57) = 104.73, p < .001, with a large effect size (partial η² = .65). The results suggest that the Social Work intervention in school can strengthen the sense of school belonging and enhances their behaviors and emotions/cognition in learning, which can be effective for students at risk of school dropout and hence could be used as a preventive intervention.