How to use Effect Size Calculator
- Choose the independent-group difference or the square of a supplied Pearson correlation.
- For independent groups, enter each sample size, mean and sample standard deviation in the same outcome units. Use SD, not standard error.
- For association mode, enter r from −1 to 1. The group summaries are ignored; this mode does not estimate correlation from observations.
- Set display precision and review the signed Cohen’s d, approximate Hedges g, or r². Keep the method and group order with the result.
Example: Effect Size Calculator
Compare group means 12 and 10 when both sample SDs are 3 and both groups contain 20 observations.
Options
- Independent groups
- Enter each sample size, mean and sample standard deviation. The difference is group 1 minus group 2.
- Association mode
- Squaring a supplied Pearson r removes its sign. It does not establish causality or a clinical interpretation.
Supported inputs and limits
- Independent groups need whole sample sizes from 2 to 1,000,000.
- Means have magnitude at most 10¹² and sample SDs run from 0 to 10¹²; both SDs zero are refused.
- Pearson r must be between −1 and 1.
- Plain decimals allow at most 64 characters, 25 digits and 12 fractional places.
- The group method uses pooled sample SD and J = 1 − 3/(4df − 1), an approximation to the Hedges correction.
More details
It is not a paired-observation method.
Display precision preserves tiny nonzero values but does not add data certainty.
There are no confidence intervals, p-values, causal conclusions or universal small/medium/large labels.
Where your input is processed
This tool does not upload your inputs. See the limits for anything saved on your device.
A standardized difference needs its model stated
The group calculation uses a pooled sample standard deviation and an approximate Hedges correction. It is for independent groups, not paired observations. Keep group order and the reported method with the number; the page does not add uncertainty intervals or judge whether a difference is practically meaningful.
Questions about Effect Size Calculator
What determines the sign of d?
Group one mean minus group two mean. Swapping the groups reverses the sign.
Is the Hedges correction exact?
No. It uses J = 1 - 3/(4df - 1), labeled as an approximation.
Does a larger effect prove causation?
No. An effect size describes these supplied summaries; study design, uncertainty and confounding remain separate.