When your data fails the normality assumption, Mann-Whitney U replaces the independent t-test — but its APA write-up trips people up because you report U and z, not t.
The template:
A Mann-Whitney U test indicated that [DV] was [higher/lower] for Group A than Group B, U = [value], z = [value], p = [value], r = [value].
What each piece means:
- Report medians, not means — the test ranks data, so means misrepresent what was actually compared.
- U is the test statistic itself; most software also gives you a normal approximation z once n is large enough, which is what you report alongside U.
- Effect size (r): divide z by the square root of total N (r = z / √N). APA 7 requires an effect size for every inferential test — this is the one reviewers ask for when it's missing.
- Common mistake: reporting only p without U/z/r. A p-value alone tells a reader nothing about magnitude.
Worked example:
A Mann-Whitney U test indicated that anxiety scores were higher for the treatment group (Mdn = 14) than the control group (Mdn = 9), U = 210, z = 2.87, p = .004, r = .41.
If you're running this test on your own data, statmate.org's Mann-Whitney calculator outputs U, z, p, and r together with this exact sentence pre-filled from your numbers, so you can skip the manual formula work.
Top comments (0)