Effect sizes with confidence intervals
Reports how large a difference or association is, with a confidence interval: Cohen's d, Hedges' g, Glass's delta, eta- and omega-squared, Cramér's V, odds ratio and risk ratio.
Menu: Analyze → Compare Means → Effect Sizes…
When should I use it?
- Every reported test — journals and APA style ask for effect sizes and their CIs.
What are the assumptions?
- Same as the matching test (e.g. normal data for d).
How do I run it in StatLab?
- Choose the design and the variables; set the confidence level.
How do I read the output?
| Output | Meaning |
|---|---|
| Cohen's d | Difference in standard-deviation units: .2 small, .5 medium, .8 large. |
| Hedges' correction | d corrected for small-sample bias. |
| Partial eta-squared (90% CI) | Share of variance explained. |
| Cramér's V, Odds Ratio | Association in a crosstab. |
Worked example
On the practice survey dataset (File → Sample dataset), this command — which the dialog's Paste button also produces — runs the procedure:
PROC esEffectSizes /design=ind /y=score /group=gender /y2='' /mu=0 /row='' /col='' /conf=95 /g1=1 /g2=2.
Output tables: Independent Samples Effect Sizes.
How would I report it in APA style?
Cohen's d = 0.10, 95% CI [−0.22, 0.42]; Hedges' g = 0.10.
Computed by StatLab's validated procedure code when this page was built. The example only illustrates the steps; choose variables that fit your own research question.
What are common mistakes?
- Calling an effect 'large' only because p is small.