I have data of glucose measurements of 100 patients taken at 2 different time points. Of these, 40 people were taking pioglitazone drug and 60 of them were on placebo. I want to test (1) whether there is change in glucose between the time points and (2) whether the change is different between drugs.

I have done a pairwise t test for my first question. But how do I get as an answer for my second question? What test should I use?

  • 1
    $\begingroup$ You're looking for mixed ANOVA. $\endgroup$
    – Eoin
    Commented Mar 26, 2023 at 9:15
  • $\begingroup$ But I have only two categories/groups. Is it fine to do ANOVA with 2 groups? $\endgroup$
    – arshad
    Commented Mar 26, 2023 at 10:01
  • 1
    $\begingroup$ Please take a look at Best practice when analysing pre-post treatment-control designs. $\endgroup$
    – dipetkov
    Commented Mar 26, 2023 at 13:33
  • 1
    $\begingroup$ You most probably want a single model that incorporates both Time and Treatment. ... Can you confirm that a single individual is measured at both Time1 and Time2 ? $\endgroup$ Commented Mar 26, 2023 at 14:55


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.