Cohen's d Calculator
Find Cohen's d and Hedges' g for two independent groups from their means, standard deviations and sizes, or from raw data.
Cohen's d
1.127
- Hedges' g (unbiased)
- 1.079
- Pooled standard deviation
- 0.8874
- Difference in means
- 1
The difference in means is 1.13 pooled standard deviations (pooled SD 0.887). The small-sample corrected effect size, Hedges' g, is 1.08.
Calculation breakdown
| Step | Formula | Result |
|---|---|---|
| Pooled standard deviation | √((10 − 1) × 0.82² + (10 − 1) × 0.95²) ÷ (10 + 10 − 2)) | 0.8874 |
| Cohen's d | (8.7 − 7.7) ÷ 0.8874 | 1.1269 |
| Hedges' g | 1.1269 × (1 − 3 ÷ (4 × 20 − 9)) | 1.0793 |
How it works
Enter each group’s mean, standard deviation and size, or switch to raw data and paste the numbers.
Worked example
The Lakens (2013) movie ratings: group 1 has mean 8.7, SD 0.82, n = 10; group 2 has mean 7.7, SD 0.95, n = 10:
- Pooled SD: √((9 × 0.6724 + 9 × 0.9025) ÷ 18) = 0.887
- Cohen’s d: (8.7 − 7.7) ÷ 0.887 = 1.13
- Correction factor: 1 − 3 ÷ 71 = 0.958, so Hedges’ g = 1.08
Rounding
Results are shown to three or four decimal places. Calculations use the unrounded values.
Sources
Frequently asked questions
What is Cohen's d?
Cohen's d is a standardized effect size for the difference between two group means. It is the difference in means divided by the pooled standard deviation, so it tells you how many standard deviations apart the groups are, whatever the units.
How do you calculate Cohen's d?
Subtract one group's mean from the other and divide by the pooled standard deviation, the square root of (n₁ − 1) times the first variance plus (n₂ − 1) times the second variance, over n₁ + n₂ − 2. In Lakens' example two groups of 10 with means 8.7 and 7.7 and SDs 0.82 and 0.95 give d = 1.13.
What is Hedges' g?
Hedges' g is a version of Cohen's d corrected for bias in small samples, g = d × (1 − 3 ÷ (4(n₁ + n₂) − 9)). In the same example g is 1.08. With large samples g and d are almost the same.
Is the sign of d important?
The sign shows which group has the larger mean. Swapping the order of the groups flips the sign and leaves the size the same. Report which group was first so readers can interpret the direction.
Does this work for paired data?
No. This calculator is for two independent groups. A paired design needs a different effect size that takes the correlation between the measurements into account.
Results are for information only. Effect sizes describe the data you enter, not whether a difference is important in practice.