Query Regarding the Use of P-Values in subsetCommunication

[kuang-da]

Hi,

Thank you for the great software; it's been incredibly useful. I have a question regarding the statistical approach used in subsetCommunication.

In the current implementation, I noticed that interaction selection is based on raw p-values. Could you please provide some insights into this design choice? Is there a specific rationale for preferring p-values over adjusted p-values in this context?

Thank you for your time.