Compare two means. Independent = two separate groups (e.g. treatment vs control). Paired = same subjects measured twice (e.g. before/after).
Mean difference—
t-statistic—
p-value—
Decision (α=0.05)—
6.0
20
Why paired is more powerful: the paired test removes between-subject noise — only the within-subject change matters. So smaller true differences become detectable than with the independent test.