Statistical Power Calculator

Find power, sample size, or effect size for one-sample and two-sample tests.

Solve for

Per group for two-sample tests

Enter the test type, tail, significance level, and the two values needed for the selected solve mode.

How this statistical power calculator works

This calculator solves for Statistical power (default), Sample size, or Effect size for a one-sample or two-sample z-test. Test type defaults to Two-sample (per group). Tail defaults to Two-tailed. Significance level α defaults to 0.05, Effect size (Cohen's d) to 0.5, and Desired power to 0.8. Sample size n starts empty, so Statistical power mode has no result until you enter n.

α must be strictly between 0 and 1. Cohen's d must be greater than 0. n must be an integer of at least 1 — a non-integer such as 64.5 is no result (this widget does not floor n). Desired power must be strictly between 0 and 1. For two-sample tests, n is per group; details then show Total sample size = 2n when you solve for sample size. The calculation is a normal (z) approximation, not a t-test.

How to use the statistical power calculator

  1. Choose Solve for: Statistical power (default), Sample size, or Effect size. Pick Two-sample (per group) or One-sample, and Two-tailed or One-tailed.
  2. Enter Significance level α strictly between 0 and 1. Enter Cohen's d greater than 0 except in Effect size mode. Enter integer Sample size n of at least 1 except in Sample size mode. Enter Desired power strictly between 0 and 1 except in Statistical power mode.
  3. Read the headline. Power is a percent. Sample size shows per group for two-sample tests, and Total sample size in the details. Effect size is the minimum detectable d.
  4. Advanced options can be opened, but it only shows Reserved for future design details. Closing it does not change the result.

Advanced options, and closing them

This section is empty of extra inputs. It only shows Reserved for future design details. Power, sample size, and effect size do not wait on Advanced options. Closing the section does not change the live result.

Power formulas

factor is 1 for One-sample and 1/√2 for Two-sample (per group). Non-centrality δ = d √n × factor. Critical z is Φ⁻¹(1 − α) for One-tailed and Φ⁻¹(1 − α/2) for Two-tailed. Power is Φ(δ − z_crit). Sample size n = ceil(((z_crit + z_power) / (d × factor))²), with z_power = Φ⁻¹(desired power). Minimum d = (z_crit + z_power) / (√n × factor).

Statistical power
power=Φ(δ−zcrit)
Required sample size
n=ceil((zcrit+zpowerd⋅factor)2)

Worked examples

Example: two-sample, two-tailed, α = 0.05, d = 0.5, n = 64

Keep Statistical power, Two-sample (per group), and Two-tailed. Enter n = 64. δ = 0.5 × √64 × (1/√2) ≈ 2.828. Critical z ≈ 1.960. Power is about 80.74%. n starts empty, so there is no result until you type it. Typed 0 for d is no result.

Example: solve for sample size at 80% power

Switch Solve for to Sample size. Desired power is already 0.8. n rounds up to 63 per group, and Total sample size is 126. The headline includes per group for two-sample tests.

Example: one-sample power with the same d and n

Switch Test type to One-sample, keep Statistical power, d = 0.5, n = 64, two-tailed α = 0.05. factor is 1, so δ = 4. Power is about 97.93%. Closing Advanced options does not change these figures. A non-integer n such as 64.5 is no result.

Things to keep in mind

This is a z approximation, not Student’s t. Two-sample n is per group. Completed results can be saved in this browser’s local history.

Statistical power calculator FAQ

Which fields are required?

Test type, Tail, and Significance level α (α strictly between 0 and 1) in every mode. Statistical power also needs Cohen's d greater than 0 and integer n of at least 1. Sample size needs d and Desired power (power strictly between 0 and 1). Effect size needs n and Desired power. Empty required fields are no result. Advanced options is optional and currently unused.

What happens when I close Advanced options?

Nothing in the result changes. The section only shows Reserved for future design details.

Is n the total sample size?

For Two-sample (per group), n is the number in each group. When you solve for Sample size, details also show Total sample size = 2n. For One-sample, n is the single sample size.

Does this use a t-test?

No. The disclaimer states that calculations use the normal (z) approximation. Two-tailed tests use z_(α/2).

Why is there no result for some n or α values?

n must be an integer of at least 1 — the widget does not floor a decimal. α and Desired power must be strictly between 0 and 1. Cohen's d must be greater than 0. In Statistical power mode, n starts empty so there is no result until you enter it.

Does this statistical power calculator store my numbers?

Completed results can be saved in this browser’s local history so you can reopen them later. They stay on your device and are not sent to our servers or to analytics.