Paired Samples t Test Calculator

When to use a paired samples t test

Use this test when the same people (or matched pairs) contribute both measurements: pretest and posttest, condition A and condition B, left and right. The test works on the difference score for each pair, which removes person-to-person variability and usually gives more power than an independent test on the same data.

If your two sets of scores come from different people, use the independent samples t test instead.

Effect size for paired designs

This calculator reports Cohen's d computed as the mean difference divided by the standard deviation of the difference scores (sometimes written d_z). It is the most common choice for paired designs, but note that it is not directly comparable to the between-groups d from an independent test.

Common mistakes

• Scores pasted in different orders, so pairs no longer match. Keep each participant in the same position in both boxes.
• Unequal numbers of scores. Every participant needs both measurements; drop incomplete cases.
• Reporting the group SDs instead of the SD of the differences when describing the test.

Paired designs go wrong quietly, usually because the pairing is not what the analysis assumes it is. If your design has repeated measures, nesting, or dropouts, a PhD statistician review is worth the hour.