Free tool that builds the APA 7 table for the Mann-Whitney U, Wilcoxon signed-rank or Kruskal-Wallis test, one row per variable: median and interquartile range per group, the statistic, z, p and the effect size SPSS does not compute. For Mann-Whitney and Wilcoxon, r = |z|/√N (Fritz, Morris & Richler, 2012); for Kruskal-Wallis, ε² = H/(N − 1) (Tomczak & Tomczak, 2014). Without z, it is approximated from U or T.
Worked example: 38 people in the control group and 41 in the intervention. Anxiety gives U = 472.5, z = −3.41, p = .001, and r = 3.41/√79 = .38, a medium effect; rumination, r = .24 (small); sleep, p = .294 and r = .12 (negligible).
Means or medians in the table? Medians and IQR: the test compares ranks, and putting means in a Mann-Whitney table is an inconsistency reviewers flag.
APA 7 nonparametric tests table with medians, IQR, U, z, p and the effect size (r = z/√N or ε²) SPSS does not compute, plus the results paragraph.