I have a matrix, the rows stand for the different diagnostic subgroups (6) and the columns stand for 4 different test statistics, making in total 24 values of test statistics. I should mention that these values are highly correlated with one another both within the diagnostic subgroup and in the different tests level. I would also like to mention that the test statistics between the different tests are in different scales. For example think of something like:

enter image description here

I have performed several tests and I want to adjust for that, given that the values on the matrix are correlated. I have also retrieves 999 more matrices under the null hypothesis. Does anyone know any procedure that adjusts for multiple testing over different tests given that the values are correlated?

  • $\begingroup$ What exactly is your scientific question? $\endgroup$ – mdewey Jul 11 '16 at 12:56
  • $\begingroup$ first I would like to know if there is any way of combining the above values (take the quadratic form for instance) so I can adjust for multiple testing among the tests (by ranking the value of the quadratic form for instance among the null mc distribution), and second if there is any relevant literature about such methods. $\endgroup$ – Akis Jul 11 '16 at 13:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.