T-Test Calculator
Run one-sample, Welch two-sample, and paired t-tests online.
Loading tool...
What is T-Test Calculator?
A t-test calculator that runs Student's one-sample, Welch two-sample, and paired hypothesis tests entirely in your browser. Paste raw measurements and instantly get the t statistic, exact p-value, degrees of freedom, Cohen's d effect size, and the 95% confidence interval of the mean difference.
Key Benefits
- No spreadsheet formulas or statistical software needed - paste values and read the verdict
- Welch correction applied automatically, matching R and scipy defaults
- Exact p-values computed numerically instead of approximated from tables
- Effect size and confidence interval included, ready for reporting standards like APA
Common Use Cases
- •Students checking lab results against a theoretical value or comparing treatment groups
- •QA engineers verifying whether a process metric drifted from its target batch mean
- •Product teams judging whether an A/B metric change is likely noise or real signal
- •Researchers running quick before/after checks on repeated measurements
How to Run a t-Test
- Choose the test type: Pick one-sample to compare against a fixed target, two-sample for independent groups, or paired for before/after measurements of the same subjects.
- Paste your data: Paste numbers into the boxes, separated by spaces, commas, semicolons, or line breaks. You can paste straight from Excel columns or CSV exports.
- Set the hypothesized mean and alpha: For a one-sample test enter the benchmark mean (mu). Keep alpha at 0.05 unless your field prescribes otherwise, then read the significance verdict.
- Read and copy the results: Check t, degrees of freedom, the p-value, Cohen’s d, and the 95% confidence interval. Copy the summary straight into your report.
Key Features
- Three test types: one-sample, independent two-sample (Welch), and paired
- Exact p-values via the regularized incomplete beta function - no lookup tables
- Cohen's d effect size plus the 95% confidence interval of the mean difference
- Smart parsing accepts spaces, commas, semicolons, and newlines; bad tokens are reported, not crashed
- Automatic Welch-Satterthwaite degrees of freedom, matching R and scipy defaults
- Copy a publication-ready summary for reports, homework, and pull requests
Frequently Asked Questions
Which t-test should I use?
Use a one-sample test when comparing a single group against a known value (e.g. does this batch weigh 500 g?). Use a two-sample test for independent groups (e.g. treatment vs control). Use a paired test when each row is a matched pair, like the same subject measured before and after.
Why are my degrees of freedom not a whole number?
For the two-sample test the calculator uses Welch-Satterthwaite degrees of freedom, which compensates for unequal variances and sample sizes. Modern tools like R and scipy use this correction by default because it keeps error rates honest when variances differ.
What does the p-value actually tell me?
The two-sided p-value is the probability of seeing a difference at least this extreme if H0 (equal means) were true. If it falls below your alpha level (commonly 0.05), the result is called statistically significant and you reject H0. A p-value is not the probability that H0 is true.
Do my data need to be normally distributed?
The t-test assumes roughly normal data, but it is robust to moderate departures, especially with 15+ values per group. For very skewed small samples consider a non-parametric alternative. Effect size (Cohen's d) around 0.2 is small, 0.5 medium, and 0.8 large.