0
$\begingroup$

I am performing several Generalized Linear Models in my analysis and I am wondering which method to use for adjusting p-values due to multiple comparisons.

I have 4 outcomes (judgement of intensity and accuracy of an emotion and 2 different muscle regions), each in 5 different models per Emotion category, (I could also combine them in one model with 5 different levels and analyze contrasts between the categories, but I decided not to do so for other reasons) and 3 different predictors

This results in 4x5x3=60 separate models

Would it be appropriate to use Benjamini-Hochberg?

Improve this question
$\endgroup$
3
  • $\begingroup$ Do you employ the same predictor observations in each model or are the predictors independent? The concern is that common data used in the various models are likely to induce strong correlations among all the p-values. $\endgroup$
    – whuber ♦
    Commented Jun 11 at 13:56
  • $\begingroup$ The centered predictors are 3 different measurements of mood. Participants fill out 3 mood questionnaires before judging stimuli and before recording muscle movements during observation of the stimuli. Hence I use the same questionnaire results as predictors for the different models. $\endgroup$
    – KayAnn
    Commented Jun 11 at 14:18
  • 1
    $\begingroup$ Benjamini-Hochberg is not a tool for "adjusting" p-values, it is a method to determine which hypotheses to reject or not given the false discovery rate you'd like to control for. $\endgroup$
    – ischmidt20
    Commented Jun 12 at 18:24

0

Reset to default