Cohen's d / Effect Size Calculator
Standardized effect size between two groups. Get Cohen's d, Hedges' g and Glass's Δ from raw data or summary statistics.
What is an effect size?
A p-value only tells you whether a difference is likely due to chance. An effect size tells you how big the difference is — and it's what you compare across studies, contexts and measurement scales. Cohen's d is the standard effect size for comparing two means.
d = (m₁ − m₂) / sₚ
sₚ = √[((n₁ − 1)s₁² + (n₂ − 1)s₂²) / (n₁ + n₂ − 2)]
Interpret d against Cohen's benchmarks (0.2 small, 0.5 medium, 0.8 large) but always in context — in a highly-controlled lab study d = 0.3 can be big; in a noisy field study it may be trivial.
Features of this calculator
- Two input modes — raw values or summary stats (means, SDs, sample sizes)
- Cohen's d, Hedges' g (small-sample corrected) and Glass's Δ side by side
- Approximate 95% confidence interval for d
- Plain-language interpretation (negligible → very large)
- Full step-by-step working (Given → Formula → Substitute → Answer)
- Handles unequal group sizes automatically
+What is Cohen's d?
Cohen's d is a standardized measure of the difference between two group means, expressed in pooled standard-deviation units. It answers 'how many standard deviations apart are these means?' independently of the raw scale of the data.
+How do I interpret d?
Cohen's rough guide: |d| ≈ 0.2 is small, 0.5 medium, 0.8 large, and 1.2+ very large. But context matters — in some fields a d of 0.1 is meaningful, in others 0.8 is expected. Always report the confidence interval and the effect size, not just p.
+What's the difference between Cohen's d and Hedges' g?
Cohen's d is slightly biased upward in small samples. Hedges' g applies a small-sample correction factor J = 1 − 3/(4(n₁+n₂) − 9). For n₁ + n₂ ≥ 50 the two are numerically almost identical.
+When should I use Glass's Δ instead?
When the two groups have very different variances (for example an intervention that changes spread as well as mean), Glass's Δ divides by the control group's SD alone instead of the pooled SD. That keeps the metric anchored to the untreated variability.
+Do I need equal group sizes?
No. The pooled-SD formula weights each group's variance by its degrees of freedom, so unequal n₁ and n₂ are handled automatically.