DataTerminology.comNEW
A standardized mean difference effect size that corrects for small sample bias, commonly used in meta-analysis.
DESCRIPTION
Hedges’ g estimates the magnitude of difference between two means, standardized by pooled variance, with a correction for small sample sizes. It enables comparison across studies with different scales. In meta-analysis, it provides a common metric for combining results and assessing practical significance. Interpretation depends on context, not just thresholds. Careful estimation and bias assessment are essential.
ALTERNATIVE DEFINITIONS
Wikipedia
Hedges’ g is a measure of effect size based on the standardized difference between two means, corrected for small sample bias.
Defines correction for small sample bias explicitly.
Cochrane Handbook
The standardized mean difference expresses the size of the intervention effect in each study relative to the variability observed in that study.
Emphasizes SMD rationale used for Hedges’ g in meta-analyses.
RELATED TERMS
Effect Size Standardized Mean Difference Meta-Analysis Glass’s Δ t-test
COMPETING TERMS
KNOWLEDGE AREAS
Mathematics Data Science Testing & Validation
TARGET AUDIENCE
Data Scientist Data Analyst Information Analyst
EXTERNAL SOURCES
- Wikipedia - Effect size#Hedges’ g
- Cochrane Handbook - SMD in meta-analysis guidance
- Statology - Applied explanation of Hedges’ g
TOP QUESTIONS
When should Hedges’ g be used?
Use Hedges’ g when comparing two groups on continuous outcomes, especially with small sample sizes or when combining studies.
Category: Requirements
How does Hedges’ g differ from Cohen’s d?
Hedges’ g applies a small-sample correction to Cohen’s d, reducing bias in effect size estimates.
Category: Comparison
What are common mistakes with Hedges’ g?
Ignoring pooled variance assumptions, misinterpreting thresholds universally, and failing to assess heterogeneity in meta-analysis.
Category: Mistakes
How is Hedges’ g interpreted?
Interpretation is context-dependent; use domain knowledge and confidence intervals rather than fixed cutoffs alone.
Category: Opinion
What risks arise when combining Hedges’ g across studies?
Publication bias, heterogeneity, and measurement differences can skew results. Use robust meta-analytic techniques.
Category: Risk
Is Hedges’ g symmetric?
Yes; swapping group labels flips the sign but not magnitude of the effect size.
Category: Definition
Can Hedges’ g be used for paired designs?
Yes, with appropriate variance and mean difference adjustments for paired or repeated measures.
Category: HowTo
How do I compute Hedges’ g?
Calculate Cohen’s d using pooled SD, then apply the small-sample correction factor J to obtain g.
Category: HowTo
Where is Hedges’ g most common?
In meta-analyses across psychology, medicine, and education to standardize effects across differing measurement scales.
Category: Example