Effect Size Calculator (Cohen's d)
Enter each group's mean, standard deviation, and sample size to calculate Cohen's d effect size.
Last updated: September 5, 2026
Frequently asked questions
What is Cohen's d?
A standardized measure of the difference between two group means, expressed in units of pooled standard deviation: d = (mean1 − mean2) ÷ pooled SD. It lets you compare effect sizes across studies that used different scales or units.
What do Cohen's d values mean in practice?
By Cohen's original conventions: about 0.2 is a small effect, 0.5 a medium effect, and 0.8 a large effect - meaning the groups' means differ by that many pooled standard deviations. These are rough benchmarks, not hard cutoffs.
How is Cohen's d different from a t-test?
A t-test tells you whether a difference is statistically significant (unlikely due to chance), while Cohen's d tells you how large that difference is in a standardized, comparable way - a result can be statistically significant with a tiny effect size, especially with large samples.
Why does sample size not appear directly in the effect size formula?
Effect size measures the magnitude of a difference independent of how much data you collected, unlike a t-test's significance, which does depend on sample size - a small, consistent difference will show a similar Cohen's d whether measured with 10 or 10,000 people, even though only the larger sample might reach statistical significance.
Can Cohen's d be negative?
Yes - a negative d just means Group 1's mean is lower than Group 2's mean. The sign shows direction; researchers usually report the absolute value when only magnitude matters.
What's an example calculation of Cohen's d?
With mean1=105, sd1=15, mean2=100, sd2=14, and equal sample sizes, the pooled SD is about 14.51, so d = (105−100)/14.51 ≈ 0.34 - a small-to-medium effect.
Is Cohen's d the same thing as a correlation coefficient?
No, though they're related and convertible. Cohen's d expresses a difference between two group means in standard-deviation units, while a correlation coefficient (like r) measures how strongly two continuous variables move together - both quantify effect size but on different scales.