Raw effect size quartiles for STEM education interventions overall are 0.21, 0.48, and 0.86 at the 25th, 50th, and 75th percentiles¶

Claim · All claims
Evidence · 1 study · quant-synthesis r3 · q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? Across 5,956 effect sizes from 95 meta-analyses, the raw absolute-Hedges'-g distribution has quartiles of 0.21, 0.48, and 0.86. → Mehmet Bicakci 2026

Evidence¶

Mehmet Bicakci 2026¶

Mehmet Bicakci, Heidrun Stoeger, Diana Wengler, Albert Ziegler. (2026). Recalibrating impact in STEM education: Empirical benchmarks for interpreting intervention effects. Educational Research Review 52. https://doi.org/10.1016/j.edurev.2026.100803

q3 · i? · quant-synthesis · r3

Umbrella review of 95 STEM meta-analyses summarizing 5,956 effect sizes; the article reports that "the 25th, 50th, and 75th percentiles were 0.21, 0.48, and 0.86" in the raw distribution. These are percentile benchmarks, not effect sizes of any single intervention, so impact is null.

"Overall, the 25th, 50th, and 75th percentiles were 0.21, 0.48, and 0.86, respectively. Subgroup heterogeneity was substantial."

Discussion¶

Related Claims¶

• Benchmarks are sensitive to precision and publication bias, but subgroup ordering remains stable and one-per-primary-study quartiles are broadly similar to raw values — related
• Cognitive, metacognitive, and strategic learning supports show the largest intervention-category benchmarks, while assessment and contextual interventions show the smallest — a narrower finding that bears on this claim
• Publication-bias adjustment via limit meta-analysis lowers the overall STEM benchmarks to 0.14, 0.32, and 0.59 — related
• STEM field benchmarks differ markedly: science shows the largest quartiles and mathematics the smallest — a narrower finding that bears on this claim