Domain-identical pretests are the most powerful single-level covariate type for student achievement, outperforming cross-domain pretests, fluid intelligence, and sociodemographic characteristics¶

Claim · All claims
Evidence · 1 study · quant-synthesis r? · q3
Currency · newest study 2024 · checked against its sources 2026-10-20 (automated, 1 of 1 entry) · not reviewed by a person

Subclaims¶

q3 i? In the meta-analysis of German large-scale assessment data, domain-identical pretests (IP) explained more single-level achievement variance than cross-domain pretests (Pooled R2 = 0.28), fluid intelligence (Pooled R2 = 0.20), and sociodemographic characteristics. → Stallasch 2024

Evidence¶

Stallasch 2024¶

Stallasch, S. E., Lüdtke, O., Artelt, C., Hedges, L. V., & Brunner, M. (2024). Single- and Multilevel Perspectives on Covariate Selection in Randomized Intervention Studies on Student Achievement. Educational Psychology Review. https://doi.org/10.1007/s10648-024-09898-7

q3 · i? · quant-synthesis · r?

Meta-analytic integration of single-level design parameters estimated from six German longitudinal samples covering grades 1-12. The article reports "Pooled R2T∣CP = 0.28 and Pooled R2T∣Gf = 0.20" for the competing cognitive covariates, with IP strongest in all grade bands.

"IP was consistently the most powerful among all covariate types: IP explained over one third of achievement differences between individual students in elementary and upper secondary school, and almost one half in lower secondary school."

Discussion¶

Related Claims¶

• At the school level, domain-identical pretests show exceptional prognostic power from grade 5 onward, but are weaker predictors at the student and classroom levels across the school career — related
• Combining domain-identical pretests with cross-domain pretests, fluid intelligence, and/or sociodemographic covariates yields only small incremental gains in explained achievement variance — related
• Sociodemographic covariates outweigh domain-identical pretests for predicting between-classroom and between-school variance in elementary school — related
• The prognostic power of domain-identical pretests degrades with longer pre-posttest time lags, most strongly at the student level — related
• General and fluid intelligence correlate strongly with mathematics achievement (r = .41 to .62), and the fluid-intelligence correlation is larger in older samples — related
• Mathematics achievement correlates more strongly with learners' knowledge, skills and cognitive abilities than with their personality traits, goals or emotions — related
• Typical R2Child values with sociodemographic and baseline covariates were around 0.18–0.26 for preschool cognitive and SEL outcomes in US data — related
• The study estimated 237 single-level, 589 two-level, and 226 three-level design parameters from German preschool probability samples — related