Jack Rossiter from the Center for Global Development Dataverse provides replication data for a 2026 study on standardised effect sizes in education. The dataset draws on 197 studies of early-grade reading interventions in low- and middle-income countries. It supports analysis of how raw learning gains relate to standardized metrics and the comparability of education program evaluations.
Use Cases
- Analyzing the relationship between raw reading fluency gains and standardized effect sizes based on the study's findings.
- Investigating the variability in effect size metrics caused by differences in test design and sample composition as described.
- Replicating the study's methodology for converting raw effects to standardized estimates across different educational contexts.
- Assessing reporting practices in education research based on the finding that fewer than one in four papers present both raw and standardized effects.
Strengths
- Based on a substantial sample of 197 education studies.
- Addresses a specific methodological issue in education research: the comparability of standardized effect sizes.
- The associated study provides concrete examples, such as a single additional word per minute corresponding to 0.03 to 0.55 standard deviations.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- The description focuses on the study's conclusions rather than detailing the dataset's specific variables and structure.
Provenance
- Source
- Center for Global Development Dataverse, author Jack Rossiter.
- Collection Method
- Likely compiled for meta-analysis or replication of the described academic study.
- Freshness
- Last updated 2026-06-01 19:39:59; freshness should be verified.
- Geography
- Focus on low- and middle-income countries, specific nations not listed.