Burnout and Teacher–Child Interactions: The Moderating Influence of SEL Interventions in Head Start Classrooms

ABSTRACT

Research Findings: The present study explored the extent to which teachers’ participation in professional development focused on children’s social-emotional learning moderated the relation between self-reported burnout and teacher–child interactions. The sample included 307 Head Start preschool teachers who participated in a large randomized controlled trial, the Head Start CARES (Classroom-based Approaches and Resources for Emotion and Social skill promotion) Project. Teachers were assigned to one of the three SEL interventions (PATHS, Incredible Years, or Tools of the Mind–Play) or a control group. Results revealed a moderating effect of treatment condition. Specifically, in control classrooms, higher self-reported burnout was related to a decline in Instructional Support scores over the course of the year. In contrast, the negative association between burnout and teacher–child interactions was not present in the intervention condition. Follow-up analyses indicated that this moderating effect was only present for teachers who were trained in the PATHS and Incredible Years interventions. Practice or Policy: Findings suggest that training and participation in interventions focused on social-emotional learning may serve as a buffer against the detrimental influence of burnout on teachers’ classroom practices.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1. In centers with five or six classrooms, all classrooms participated in the study. In centers with more than six classrooms, five classrooms were randomly selected.

2. Two of the three dichotomous variables containing the teachers’ years of teaching experience (i.e., < 3 years, 3 to <10 years), had to be dropped from the publically available data set because the frequencies of values were too low to be masked. A variable providing information on whether teachers have less than 10 years of experience (1 = yes, 0 = no) was maintained in the dataset.