I tried to use Mann-whitney test, but when i read the assumptions it shouldn't be overlap between the samples. Should I use Wilcoxon test ? But they are not the same study samples as well
The problem here is that if the groups really do overlap then the samples are not independent. Statistics pretty much operates under the assumption thatthe groups are independent of each other .if they aren't then you have only one group and hence only one mean say for the whole group and hence nothing to test. however you could look at classification methods and ask questions like how much overlap do you have between the subgroups and what does that mean? The attachment contains information about my favorite cluster analysis book, you can download it from the z-library. The computer programs are available in R package cluster. Google the CRAN website for details about getting them. Best wishes David Booth
The t-test is commonly used in statistical analysis. It is an appropriate method for comparing two groups of continuous data which are both normally distributed. The most commonly used forms of the t- test are the test of hypothesis, the single-sample, paired t-test, and the two-sample, unpaired t-test.
Assuming that you mean some of the people in Group 1 are also in Group 2, do you have person identifiers so that you can link those people in both samples?