Word-Problem Interventions: A Meta-Analysis of 65 Years of Research
Vessonen, T., et al. · 2025
grade Cmeta-analysisindependentreplicated
Sample
114 studies / 531 effect sizes / N=20,456
Population
K-12.
Design
Modern statistically competent replacement for Peltier & Vannest 2017 (whose g=1.57 with near-zero-width CIs reflected improper pooling).
Key findings
Word-problem intervention is among the most robustly positive domain-skill literatures in math — with the aligned-measure inflation confirmed inside the meta itself.
Genetic confound
Experimental base.
Replication notes
Confirms schema-instruction direction; the standardized-measure effect is the policy-relevant number and is substantially below the headline.
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Word-problem interventions, overall | g | 0.95 | mixed | immediate | unclear | end-of-treatment | domain-skill |
| Measure-type moderator | contrast | researcher-developed measures significantly larger than standardized — inflation confirmed within the meta | mixed | immediate | unclear | end-of-treatment | domain-skill |
Cited by
- Word problems, schema instruction, and the conceptual-vs-procedural questionmoderate supportconf: highgc: low