Based on Cohen (1988)
Select Statistical Test
t-Test for Means
One-sample, two-sample (equal or unequal sizes), or paired t-test.
Solve For
Cohen's Effect Size Presets
Effect Size Reference
Cohen's conventional benchmarks
TestES MeasureSmallMediumLarge
t-testd0.200.500.80
Proportionh0.200.500.80
Correlationr0.100.300.50
ANOVAf0.100.250.40
Chi-squaredw0.100.300.50
Regression0.020.150.35
Power (1−β)InterpretationCommon Use
0.50UnderpoweredExploratory
0.80StandardMost research
0.90High powerClinical trials
0.95Very highHigh-stakes decisions
Alpha (α)Type I Error RiskCommon Use
0.1010%Exploratory studies
0.055%Standard convention
0.011%Strict criteria
0.0010.1%Genome-wide studies
Minimum N required per test (power = 0.80, α = 0.05, two-sided) at Cohen's conventional effect sizes.
📊 t-Test (Two-Sample)
Small (d=0.2): 394 per group
Medium (d=0.5): 64 per group
Large (d=0.8): 26 per group
pwr.t.test
n per group × 2 = total N
📊 t-Test (One-Sample / Paired)
Small (d=0.2): 197
Medium (d=0.5): 34
Large (d=0.8): 15
pwr.t.test
type = "one.sample" or "paired"
🎯 One Proportion
Small (h=0.2): 197
Medium (h=0.5): 33
Large (h=0.8): 14
pwr.p.test
Uses arcsine transformation
🔀 Two Proportions (Equal n)
Small (h=0.2): 394 per group
Medium (h=0.5): 65 per group
Large (h=0.8): 26 per group
pwr.2p.test
h = ES.h(p1, p2)
🔗 Pearson Correlation
Small (r=0.1): 782
Medium (r=0.3): 84
Large (r=0.5): 28
pwr.r.test
Fisher Z' transformation applied
📈 One-Way ANOVA (k=4 groups)
Small (f=0.1): 274 per group
Medium (f=0.25): 45 per group
Large (f=0.4): 18 per group
pwr.anova.test
Balanced design (equal group sizes)
χ² Chi-Squared (df=2)
Small (w=0.1): 964 total
Medium (w=0.3): 108 total
Large (w=0.5): 40 total
pwr.chisq.test
df = (rows−1)(cols−1)
📐 Regression / GLM (u=5)
Small (f²=0.02): N≈485
Medium (f²=0.15): N≈92
Large (f²=0.35): N≈50
pwr.f2.test
N = v + u + 1; u = predictors
📊

Select a test, fill in parameters,
and click Calculate to see results.

Result
Sample Size
Statistical Power

🎯 What is Power?

Statistical power (1−β) is the probability of correctly rejecting a false null hypothesis. The conventional minimum is 0.80 (80%).

⚠️ Type I vs Type II

α = False positive rate (Type I error). β = False negative rate (Type II). Power = 1−β.

📐 Effect Size

Standardized measure of the magnitude of an effect. Larger effects require smaller samples to detect at the same power level.

📚 Reference

All calculations follow Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences, 2nd ed.