Free tool

P-Value & t-Test Calculator

Run a two-sample or one-sample t-test from summary statistics, or convert a z or t statistic straight to a p-value.

Two-sided p-value
0.0083

t = -2.7937, df = 36.25, mean difference = -2 (SE 0.7159)

Cohen's d = -0.883 · Below 0.05 at the conventional threshold

Built by Quandra Sciences

We build verifiable AI for regulated biotech work. These bench tools are free — no account, no email, no tracking wall.

  • Deterministic math, shown in the open
  • Runs entirely in your browser
  • No sign-up required
Reserve founder pricing

Method notes

Assumes independent observations and approximately normal sampling distributions. Paired designs, multiplicity adjustment, and non-normal endpoints need a pre-specified analysis plan — this tool does not substitute for one.

Using p-values responsibly

Welch vs. pooled

Welch's t-test does not assume equal variances and is the safer default. Use the pooled version only when equal variance is justified and pre-specified.

Report the effect, not just p

A p-value tells you about compatibility with the null, not magnitude. Always pair it with the mean difference and a confidence interval.

Multiplicity

Testing many endpoints inflates false positives. Regulated submissions require a pre-specified multiplicity strategy such as hierarchical testing or Holm adjustment.

Underpowered studies

A non-significant result in a small study is not evidence of no effect. Check the confidence interval width before concluding anything.

What Quandra v8 adds

These calculators give you one number at a time. Quandra v8 writes the statistical analysis plan, generates R/SAS code, cross-checks assumptions against a locked reference corpus, and keeps an audit trail for regulated work.