I would like to check the proportionality assumption of a proportional hazards model. The problem is that I am doing multiple imputation.
I have checked the assumption in every one of the imputed data set. If the assumption is violated in only one imputation, does it means the assumption is rejected?
Is it reasonable to apply something like Bonferroni correction with the number of imputed data sets?
If possible, I would like to have a reference for any given suggestion.