Hedges' g measures the standardized difference between two group means while correcting Cohen's d for potential bias from small record counts. Like Cohen's d, it compares the mean difference with pooled variability; unlike a significance test, it describes effect magnitude rather than deciding whether to reject a null hypothesis. section 13.8.3 section 20.3
Use Hedges' g to interpret continuous-data comparisons:
- Prefer it when groups contain fewer than 20 records, where Cohen's d may provide an inflated standardized difference. section 13.8.3
- Read values near 0.2 or below as small, around 0.5 as medium, and 0.8 or above as large. section 20.3
- Keep the sign: reversing the group arguments reverses the estimate, while the magnitude describes the standardized difference. section 13.8.3
- Pair it with a t-test rather than using it as a replacement, because effect sizes do not account for record counts in the same way significance tests do. section 13.8.3 section 20.3
Comparing Gini-index groups in R
Run hedges_g(ws_giniX$gini_index, ws_giniY$gini_index) from the effectsize package. The result is -0.32 with a 95% confidence interval from -0.70 to 0.07, indicating a small standardized effect; swapping the arguments produces 0.32. section 13.8.3
cohens_d(ws_giniX$gini_index, ws_giniY$gini_index) ## Cohen's d | 95% CI ## ------------------------- ## -0.32 | [-0.70, 0.07] ## ## - Estimated using pooled SD. hedges_g(ws_giniX$gini_index, ws_giniY$gini_index) ## Hedges' g | 95% CI ## ------------------------- ## -0.32 | [-0.70, 0.07] ## ## - Estimated using pooled SD. glass_delta(ws_giniX$gini_index, ws_giniY$gini_index) ## Glass' delta | 95% CI ## ----------------------------- ## -0.51 | [-0.93, -0.09]
Compare Cohen's d, Hedges' g, and Glass's delta outputs side by side, including confidence intervals and pooled-standard-deviation estimates. section 13.8.3
