Paired T-Test Calculator

Use the paired t-test when the same people are measured twice (before and after, two conditions) or when observations come in matched pairs. Enter the two columns in the same order, or just the mean and SD of the differences. You get t, df, p, the confidence interval of the mean difference, Cohen’s dz and an APA sentence.

When to use a paired t-test

Pairing removes the differences between individuals from the comparison, so a paired test is usually far more sensitive than an independent test on the same numbers. Typical designs:

  • Before-and-after measurements on the same participants.
  • Two treatments given to each participant in turn (a crossover design).
  • Matched pairs, such as twins or patients matched on age and sex.
  • Two measurement methods applied to the same items.

The test assumes the differences are roughly normally distributed. The raw scores don’t need to be. If the differences are clearly skewed or contain outliers, use the Wilcoxon signed-rank test instead.

Formula

d = x₁ − x₂ for each pair · t = d̄ ÷ (s_d ÷ √n) · df = n − 1 · d_z = d̄ ÷ s_d

Worked example

Eight students scored 85, 90, 78, 92, 88, 76, 95, 89 before a revision course and 80, 82, 75, 85, 79, 70, 88, 84 on a parallel test afterwards. The mean difference is 6.25 points with an SD of 1.91, so t(7) = 9.26 and p < .001. Treating the same numbers as two independent groups gives t = 2.01 and p = .064. That’s why it matters to pick the right test.

Frequently asked questions

What if some pairs have missing values?

Remove the incomplete pairs before pasting. Both lists must contain the same number of values in the same order, and the calculator will tell you if they don’t.

Which effect size should I report?

Cohen’s dz (mean difference ÷ SD of the differences) is the standard effect size for paired designs and the one G*Power uses for sample-size planning.