T-Test Calculator
One-sample, two-sample (Welch or pooled), and paired t-tests with APA write-up.
About this tool
Paste one or two samples and pick one-sample, independent two-sample (Welch or pooled) or paired. You get t, df, p, the mean difference, a confidence interval and Cohen's d.
How to use
How it works
Welch's test does not assume equal variances and uses a fractional df. Paired mode tests the mean of the differences between matched observations. The p-value is the chance of a t at least this extreme if the null were true; the confidence interval shows the plausible range for the mean difference.
Welch: t = (x̄₁ − x̄₂) ÷ √(s₁²/n₁ + s₂²/n₂) one-sample: t = (x̄ − μ₀) ÷ (s/√n)
Worked example
Two groups of eight scores
- Input
- A: 88,92,79,95,85,90,83,97 B: 81,86,78,90,80,84,76,89 (Welch, two-tailed)
- Output
- t(13.56) = 2.00, p = .066; mean difference 5.625; 95% CI [−0.43, 11.68]; d = 1.00
- Walkthrough
- Means 88.625 and 83.0; variances 37.41 and 26.0; SE = √(37.41/8 + 26/8) = 2.815; t = 5.625 ÷ 2.815 = 1.998. The CI spans zero, matching p >.05.
Limitations
- Needs raw data, not summary statistics.
- Assumes roughly normal samples; very small or skewed samples call for a non-parametric test.
Questions
When should I use Welch instead of pooled?
What is the paired t for before-and-after 70,75,80,68,72 vs 72,78,83,70,75?
What is the one-sample t for 5.1,4.9,5.6,5.8,6.0,5.2 against 5?
Does a p of.066 mean no difference?
Features
- One, two and paired
- Confidence interval
- Cohen's d
Guides
Related resources
Guides and videos from around the web—opens on YouTube or the publisher’s site.
Finding mean, median, and mode
Khan Academy
Baseline summary stats for coursework checkoffs.
Watch on YouTubeMean, Variance and Standard Deviation
StatQuest with Josh Starmer
Descriptive stats behind table and test tools.
Watch on YouTube
- Statistics and probability
Khan Academy
Full intro stats path next to hub calculators.
- Student's t-distribution
Wikipedia
Distribution reference behind the t-table tool.