Distinguish subgroup stratified analyses from statistical moderation tests when evaluating differential SEL effects¶

Strategy · All strategies
Evidence · 2 claims (2 for) · 2 studies (2 review), q2 · 0 of 2 report an effect size · 2 claims rest on one study

Description¶

The review advises that depending on research aims, subgroup analyses or moderation tests of diversity characteristics may be appropriate, and that data may direct the decision, for example when subgroups have different variance-covariance structures. It cautions that stratified analyses do not statistically compare groups, so heterogeneity claims should not rest on separate within-group tests. Both test types were included in the review's moderation category while recognizing they differ in interpretation.

Design Implications¶

Context¶

Requirements¶

• Analysts must recognize that stratified subgroup analyses and formal moderation tests should not be interpreted the same way

Constraints¶

• Three-way interaction moderation tests were excluded from the review because unique moderator effects could not be separated

Target Learners¶

• children and adolescents in universal school-based SEL programs

Target Learning Goals¶

• accurate evaluation of differential program effects across student groups

Claims¶

• Most reviewed SEL studies did not test diversity characteristics as moderators, and those that did found inconsistent moderation effects [+W]
• Program and delivery characteristics moderate SEL effects: implementation quality, duration, skill practice, and delivery agent [+W]

Related Strategies¶

Examples¶

-

Key Sources¶

• Rowe, H. L., & Trickett, E. J. (2018). Student Diversity Representation and Reporting in Universal School-Based Social and Emotional Learning Programs: Implications for Generalizability. Educational Psychology Review. https://doi.org/10.1007/s10648-017-9425-3