I am running a Wilcoxon Sign-Rank test with two lists. One contains 5 elements and the other contains 6. They were taken from the same place but under different conditions. I am trying to compute the test results, but I understand that I need each list to have the same number of elements. How should I address this issue? I was considering representing the mission value as a NaN element.
1
-
2$\begingroup$ You are confusing two different things that are easily confused because they are similarly named but are not the same. The Wilcoxon signed-rank test is for paired data with the same number of occurrences in two groups, that you could think of as a before and after test on the same subjects, and the Wilcoxon rank-sum test more often called the Mann–Whitney U test for unpaired data that can have different numbers of observations in each group. $\endgroup$– CarlCommented Mar 29 at 4:24
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
The Wilcoxon signed rank test requires pairs of observations, often before-after pairs.
Are you observations paired or matched? If not, you've chosen the wrong test.
answered Mar 29 at 3:42