0

I have two groups that are particularly unequal 111 vs 29 persons in each group and i would like to know if the Mann-Whitney U test works fine in this case also. I know from previous discussions that when the two groups are not similar 1:2 ratio this works fine. But what about an almost 1:4 allocation?

Pergialiotis Vasilios
  • 1
  • A number of other answers on site deal with a variety of large ratios of sample size. It doesn't matter if it's a million vs 3. – Glen_b Aug 10 '17 at 11:57

1 Answers1

1

It does not really matter. With more even allocation you would have more power to reject the null hypothesis, if there is in truth a tendency to have higher/lower values in one of the groups compared to the other. Otherwise there is no issue.

Björn
  • 21,227
  • 2
  • 26
  • 65