Ask Mallard

Paired t-test

§2.5
Rev 4.6 · reviewed Feb 2026
Paired t-test
For before/after measurements on the same patients. Work with the differences: enter the number of pairs, the mean difference, and the SD of the differences.
Source formula: Student. Biometrika 1908;6:1–25.
Table. Paired differences (after − before)
Pairs (n)Mean differenceSD of differences
* Student. Biometrika 1908;6:1–25.
Mean difference (95% CI)
-4.60 (-7.70 to -1.50)
copy
t statistic (df)
-2.91 (24)
copy
Two-sided p
0.008▲
copy
On average, values changed by -4.60 units within patients. The change is statistically significant (p 0.008). Pairing removes between-patient variability, which is why this design is efficient.
Understand what this is measuring