Ranstam Review DataBase (RR.db)

Templates for phenomenon "testing"

  1. It is described in the statistics section that "Means were compared using Student’s t-test. Medians were compared using the Wilcoxon rank-sum test. Proportions were compared using the Fisher’s exact test. Kaplan-Meier survival and time to event analyses were performed with log-rank p-value testing." This is not formally correct because the test functions are used to test hypotheses about the populations represented by the studied sample, not the observed means, medians and proportions. Furthermore, Wilcoxon's Rank-Sum Test is based on the null hypothesis of 50% probability that a random value from one population exceeds a random value from the other population. A null hypothesis of identical medians can only be tested with the additional assumption that the shapes of the distributions are identical in the two populations.


[Return]