t-test calculator

Check whether the average of a metric differs between two independent groups. The calculator uses Welch's t-test, which does not assume equal variances.

Welch's t-test — a statistical test for comparing the means of two independent groups whose variances may differ.

Calculation FAQ

Group values

Group A

Numbers separated by commas or new lines

Group B

Numbers separated by commas or new lines

t-statistic

—

Result

Welch's t-test

Interpretation

Change the inputs to recalculate.

In SurveyNinja, NPS, CSAT and more are calculated automatically. Create a survey in 5 minutes and get real-time analytics.

Start free

Data is not stored — calculation runs in your browser.

How Welch's t-test is calculated

Formula

t = (meanA − meanB) / sqrt(sdA²/nA + sdB²/nB). Degrees of freedom follow the Welch–Satterthwaite formula.

Means

The test compares the difference in means with the spread inside the groups.

Welch

Works when group sizes and variances differ. It is the safer default.

p-value

Shows how compatible the observed difference is with the hypothesis that the means are equal.

Calculation examples

1 Interface A/B rating

A: 7, 8, 8, 6, 9

B: 8, 9, 9, 7, 10

If the p-value is below 0.05, the difference in means can be considered statistically significant.

2 Task completion time

Old version: 55, 60, 52

New version: 43, 47, 41

The test shows whether there is enough data to conclude that the new version is faster.

What to avoid

Using it for paired measurements If the same people went through both conditions, you need a paired t-test, not an independent Welch's t-test.
Ignoring the shape of the distribution With very small samples and strong outliers, the result can be unstable.
Confusing significance with effect The p-value does not show how much the difference matters for the business. Look at the effect size too.

Frequently asked questions

Why Welch's t-test?

It is more robust than the classic Student's t-test because it does not require equal variances or equal group sizes.

What does the p-value mean?

The probability of seeing a difference this large or larger if the means were actually equal.

Can I use it for a 1–5 scale?

In practice people do, when they treat the scale as interval. For rigorous analysis, add a non-parametric test.

How many observations do I need?

The smaller the effect and the larger the spread, the bigger the sample must be. With very small groups, draw conclusions carefully.

What if the p-value is above 0.05?

It does not prove there is no effect. The effect may be small or the sample too small; check the power and the effect size.

All calculators

Automate your metrics