Cohen’s d Calculator

Enter the mean, standard deviation and size of two groups to get Cohen’s d, the standardized difference between them, with an exact confidence interval. The calculator also gives Hedges’ g, works from a reported t value or a paired design, and turns d into a probability that readers can understand.

Start from

Two independent groups.

Group 1 (treatment)

Group 2 (control)

Confidence level

Load a worked example

Cohen’s d

0.5095% CI −0.10 to 1.10

A medium effect by Cohen’s benchmarks: the group means differ by 0.50 standard deviations, but the interval includes 0.

Cohen’s d
0.501
95% CI −0.099 to 1.096
Hedges’ g (bias-corrected)
0.492
95% CI −0.097 to 1.076
Glass’s Δ (control SD)
0.421
Pooled SD
1.5967
Standard error of d
0.3046
df = 43
Correlation r (point-biserial)
0.243
equal group sizes
r² (variance explained)
5.9%
Odds ratio equivalent
2.48
exp(dπ/√3)
Probability of superiority
63.8%
common-language effect size
Cohen’s U3
69.2%
group 1 above the group 2 mean
Overlap of distributions
80.2%
NNT (Kraemer)
3.6

Report it like this

Group 1 (M = 5.2, SD = 1.1, n = 20) scored higher than group 2 (M = 4.4, SD = 1.9, n = 25), Cohen’s d = 0.50, 95% CI −0.10 to 1.10 (Hedges’ g = 0.49).

How to use the calculator

  1. Enter both groups. Put the treatment or experimental group first. A positive d means group 1 scored higher.
  2. Choose the confidence level. 95% is standard.
  3. Read d and its interval. The interval shows the range of population effects compatible with your data.
  4. Copy the sentence for your results section.

If a paper reports only t and the group sizes, switch to “t statistic”. For a within-subject design, choose “Paired” to get dz.

Cohen’s d formula

d = (M₁ − M₂) / SDpooled,   SDpooled = √[((n₁ − 1)SD₁² + (n₂ − 1)SD₂²) / (n₁ + n₂ − 2)]

The pooled SD weights each group’s variance by its degrees of freedom. When the groups are the same size it is simply the root mean square of the two SDs. Some textbooks pool with n instead of n − 1; the difference is negligible except in very small samples.

d = t √(1/n₁ + 1/n₂),   dz = t / √n

The confidence interval is exact: it inverts the noncentral t distribution that the observed t follows, which is why it is not symmetric around d.

Worked example

A treatment group of 20 scored a mean of 5.2 (SD 1.1) and a control group of 25 scored 4.4 (SD 1.9), the default values.

  • Pooled SD = √[(19 × 1.1² + 24 × 1.9²) / 43] = √(109.63 / 43) = 1.597.
  • d = (5.2 − 4.4) / 1.597 = 0.50, a medium effect by Cohen’s benchmarks.
  • 95% CI −0.10 to 1.10: the data are compatible with anything from a tiny negative effect to a large positive one.
  • Hedges’ g = 0.50 × (1 − 3/171) = 0.49.
  • Probability of superiority = 64%: a randomly chosen treated person scores higher than a randomly chosen control 64% of the time.

The matching Welch t-test gives p = .085, consistent with an interval for d that includes 0.

Interpreting Cohen’s d

  • d = 0.2 (small): the groups overlap by 92%; a treated person beats a control 56% of the time.
  • d = 0.5 (medium): 80% overlap; 64% probability of superiority. Cohen described it as visible to the naked eye.
  • d = 0.8 (large): 69% overlap; 71% probability of superiority.

These labels are rough guides from the behavioural sciences, not rules. Judge an effect against what is typical in your field and what difference would matter in practice. To plan a study that can detect a given d, use the power analysis calculator; to convert d to r or an odds ratio, see the effect size calculator.

How to report Cohen’s d

APA style asks for an effect size with every test and recommends a confidence interval for it. Give d to two decimals, with the means, SDs and sample sizes it came from:

The treatment group scored higher (M = 5.20, SD = 1.10, n = 20) than the control group (M = 4.40, SD = 1.90, n = 25), t(39.55) = 1.77, p = .085, Cohen’s d = 0.50, 95% CI −0.10 to 1.10.

Say which SD you standardized by. For paired data, write dz and explain that it uses the SD of the differences.

Questions

How do I calculate Cohen’s d?
Subtract the mean of group 2 from the mean of group 1 and divide by the pooled standard deviation. For means of 5.2 and 4.4 with SDs of 1.1 and 1.9 (n = 20 and 25), the pooled SD is 1.597 and d = 0.8 ÷ 1.597 = 0.50.
What is a good Cohen’s d?
Cohen suggested 0.2 as small, 0.5 as medium and 0.8 as large. Whether an effect is important depends on the outcome and the cost of the intervention: a d of 0.2 on mortality can matter more than a d of 0.8 on a questionnaire score.
Can Cohen’s d be negative?
Yes. A negative d means group 2 has the higher mean. The sign depends only on which group you enter first, so report the direction in words as well.
Can Cohen’s d be greater than 1?
Yes. d has no upper limit. A d of 1.5 means the means are one and a half standard deviations apart. Very large values are common for paired designs (dz) and for manipulation checks.
How do I calculate Cohen’s d for a paired t-test?
Divide the mean of the differences by the standard deviation of the differences, or divide t by the square root of the number of pairs. This is called dz. For t(7) = 3.99 from 8 pairs, dz = 3.99 ÷ √8 = 1.41.
How is the confidence interval for Cohen’s d calculated?
The calculator inverts the noncentral t distribution: it finds the population effects for which the observed t would fall at the 2.5th and 97.5th percentiles. This exact method is preferred over the normal approximation, especially in small samples.

Sources