Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It only takes a minute to sign up.
Sign up to join this community
Stats newbie here currently trying to teach myself some stats topics. When testing for evidence that mean earnings per share were higher in 1988, should I use Wilcoxon Signed Rank since the samples can be paired?
But I'm still confused about the different uses of tests. What changes in the details of the problem would be required for me to be able to use Mann Whitney U, a t-test, and a z-test?
Thank you!
The data values are paired so you would use the Wilcoxon signed rank or paired t-test. An example of unpaired values might be if you had ten corporations from the United States and you wanted to compare them with ten corporations from Japan. In the unpaired case of course the sample sizes do not have to be equal.
self-studyquestion. You should theself-studytag, read its tag wiki and follow the guidelines there (which you partly do already but you should explain your thoughts/understanding of the question in more detail if possible). On that question -- it's not that the samples can be paired -- they are paired. That should not be ignored. $\:$ 2. The question at the end is not a self-study question and should be asked separately. $\endgroup$