0
$\begingroup$

In a study where I compare the male vs. female university students (independent variable = gender) on 20 different outcomes (each referring to a wellbeing measure), can I apply a false discovery rate (FDR)? Across the literature it seems that FDR is predominantly used for situations where your independent variable has more than two levels. However, this experiment conducts multiple tests using the same individuals, which I interpret to mean that there is familywise error that needs to be accounted for. Am I correct in my assessment or the fact that my independent variable has only two groups means it is not necessary to correct for type I error, as there is only one pairwise comparison for each dependent variable.

$\endgroup$

1 Answer 1

1
$\begingroup$

As far as I know, FDR corrections make no requirements on the number of levels of your variables. Multiple comparison corrections are simply ways to account for the fact that type I errors occur and accumulate as you make more comparisons.

In general, if you have a distribution of p-values from which you're making hypothesis tests, you can perform a multiple comparison correction.

For example, the Bonferroni, a simple, very conservative correction, doesn't make any assumptions about independence and can just be used as is.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.