A practical guide to causal moderation analysis for investigating the role of context, identity, and culture in intervention research

• Published in: Journal of school psychology Vol. 111; p. 101473
• Main Authors: Dong, Nianbo, Herman, Keith, Kelcey, Benjamin, Ren, Sirui, Reinke, Wendy, Spybrook, Jessaca
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.08.2025
• Subjects: Average moderated treatment effect across the entire sample (AMTE); Average moderated treatment effect on targeted subgroups (AMTS); Child; Culture; Generalized propensity scores; Humans; Inverse probability treatment weighting; Potential outcome; Psychology, Educational - methods; Schools; Social Identification; Students - psychology; Generalized propensity scores; Inverse probability treatment weighting; Average moderated treatment effect on targeted subgroups (AMTS); Potential outcome; Average moderated treatment effect across the entire sample (AMTE)
• ISSN: 0022-4405; 1873-3506; 1873-3506
• DOI: 10.1016/j.jsp.2025.101473

Contextual, identity, and cultural factors are not only associated with student outcomes but can also serve to moderate the effects of interventions. However, the conventional analysis of moderation commonly used in school psychology is subject to the selection bias potentially introducing bias into estimated moderator effects. This article provides a tutorial on the application of a recently developed framework for causal moderation effects within the context of investigating contextual, identity, and cultural effects in intervention research. Using an example that focuses on the effectiveness of a teacher classroom management program in improving students' social and behavioral outcomes, we demonstrate the application of the (stabilized) Inverse Probability Treatment Weighting method to investigate heterogeneous treatment effects. The analysis aims to scrutinize not only the Average Moderated Treatment Effect (AMTE) across the whole sample but also the Average Moderated Treatment Effects on Targeted Subgroups (AMTS). Finally, we discuss our findings and conclude with suggestions for the application of causal moderation analysis.