Menu

Free Statistics Calculator

t-test Calculator

Compare two groups and get an exact p-value in seconds. Paste your numbers, and the calculator picks the right test - Student's t-test, Welch's t-test, or the Mann-Whitney U test - then hands you a paste-ready methods sentence for your paper.

No sign-up requiredRuns on scipy - exact, reproducibleFree to use

Enter your groups

0 values

0 values

Test (two groups)
Error bars

How to compare two groups

A two-sample (unpaired) t-test asks whether the means of two independent groups differ by more than you would expect from random sampling. Enter each group's measurements, choose Auto-pick, and the calculator validates the test's assumptions for you before reporting the p-value.

A p-value below 0.05 is the conventional threshold for "statistically significant," meaning the observed difference is unlikely under the null hypothesis of no difference. The calculator also reports significance stars (* p<0.05, ** p<0.01, *** p<0.001) so the result drops straight onto a figure.

When to use the Mann-Whitney U test

When a group is small or clearly non-normal, the parametric t-test can mislead. In that case the Mann-Whitney U test compares the distributions using ranks instead of means and makes no normality assumption. Auto-pick switches to it automatically when the Shapiro-Wilk check fails. Comparing three or more groups instead? Use the ANOVA calculator.

Frequently asked questions

Which t-test does this use?

On Auto-pick, the calculator checks normality (Shapiro-Wilk) and equal variance (Levene). If both groups are approximately normal with equal variances it runs a Student's t-test; if variances differ it runs Welch's t-test; if a group is not normal (or too small to judge) it runs the non-parametric Mann-Whitney U test. You can also force a specific test.

What is the difference between Student's and Welch's t-test?

Both compare two means, but Student's t-test assumes the two groups have equal variances while Welch's t-test does not. Welch's is the safer default when group sizes or spreads differ, which is why the auto-picker selects it when Levene's test flags unequal variance.

Is my data sent anywhere or stored?

Your numbers are sent to Plotivy's compute service to run the test with scipy and are not stored. Nothing is saved to an account because no sign-up is required.

How do I report the result in a paper?

Copy the generated methods sentence, which names the test, the per-group sample sizes, and the error-bar convention. Pair it with the p-value and the group means shown in the descriptives table.

Related tools

From numbers to a figure

Plotivy runs the test, picks the right chart, and annotates significance bars on a publication-ready figure - all from your data.

Open the Analyze tool