StatReasonHypothesis Tests
A hypothesis test asks whether your data is consistent with a specific stated null model, or provides evidence against it at a chosen significance threshold. Every test here renders its null hypothesis, selected alternative, alpha, test statistic, degrees of freedom where relevant, p-value, and an effect size or confidence interval together, since a p-value in isolation cannot tell you how large or meaningful an effect actually is.
These tools calculate sample size, ties, and expected counts directly from your data, but the design questions, whether your groups were validly sampled, whether the comparison is fair, whether independence holds, are left for you to verify, since no test statistic can establish those from numbers alone.
Engines in this category
- Independent t-Test
Compare the means of two independent groups using Welch’s t-test by default.
- Chi-Square Test of Independence
Test whether two categorical variables are associated in a contingency table.
- Paired t-Test
Compare two measurements taken on the same subjects or matched pairs.
- Two-Proportion z-Test
Compare two independent proportions, such as two conversion or event rates.
- One-Sample t-Test
Compare a single sample mean against a fixed benchmark value.
- Levene Variance Test
Test whether two or more groups show evidence of unequal variance.
- One-Proportion z-Test
Test a single observed proportion against a fixed benchmark rate.
- Chi-Square Goodness-of-Fit Test
Test whether observed category counts match a stated expected distribution.
- P-Value Calculator
Convert a statistic and named null distribution into a tail probability.
- Exact Binomial Test
Test a binomial proportion exactly, with no normal-approximation assumption.
- Fisher's Exact Test
Test 2 by 2 association exactly, safe for sparse expected counts.
- McNemar Test
Compare paired binary outcomes using only the discordant pairs.