For paired data the interval is built from the differences within each pair. Enter both columns in the same order, or the mean and SD of the differences.
d̄ ± t*(n − 1) × s_d ÷ √n
Example
25 patients’ blood pressure falls by an average of 4.2 mmHg (SD of the differences 5.1). The 95% interval is 4.2 ± 2.064 × 5.1 ÷ √25 = [2.09, 6.31] mmHg. Because the interval excludes 0, the reduction is statistically significant, and it is plausibly between about 2 and 6 mmHg.
The interval matches the paired t-test: if the 95% CI excludes 0, the two-tailed test is significant at α = 0.05. The interval also shows the size of the change, which the p-value alone cannot.