Using the car package, I drew two qqPlots of 144 p values from correlated tests on one dataset. The top plot uses a reference line drawn by "quartiles" and the bottom plot uses a line drawn by "robust" regression.

This is concerning as the two seem to be saying different things about my data. I am tempted to go with the robust regression because after correcting for multiple comparisons, I obtain 1/144 results < 0.05 after Bonferonni and 6/144 results < 0.05 after FDR, which seems more consistent with the second graph.

However, I also have some concerns about my study being underpowered and this paper (https://folk.uio.no/tores/Publications_files/Schweder_Spjotvoll_1982.pdf) suggests that the first plot could be an indicator of that. Which plot would be better in my case? I would be happy to provide additional detail if needed.

quartiles robust


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.