Paired measurement (dz, dav)

In a paired measurement, before and after come from the same people. The test compares the two values for each person. dz relates the mean change to the spread of the changes (dz = mean of the differences divided by their standard deviation); dav instead divides by the mean of the two standard deviations of before and after and is comparable with d from group comparisons. That is why the classification small, medium, large follows dav; dz serves statistical power.

Example: Knowledge test before and after the course, 14 people: on average 2.2 points more, dz = 1.36, dav = 0.77. The change is medium-sized: dav is just below the benchmark for a large effect (0.8).

Caution: Without a comparison group the measurement only shows that something changed - not why. Practice, time and other influences are possible. Cohen's benchmarks (0.2 / 0.5 / 0.8) apply to d and dav, not to dz: dz is larger the more strongly before and after are correlated (Lakens 2013).