p-values for multiple AUCs

[Sara Hansen]

I have a data with a ground truth (truth) variable and 4 different groups each resulting in a AUC. See dataexample below:

* Example generated by -dataex-. For more info, type help dataex
clear
input byte(id truth group1 group2 group3 group_comb)
 1 0 0 0 0 0 
 2 1 1 1 1 1 
 3 0 1 0 1 0 
 4 0 0 0 0 0 
 5 1 0 1 1 0 
 6 1 1 1 1 1 
 7 0 1 0 0 1 
 8 0 0 0 0 0 
 9 0 0 0 1 0 
10 1 0 0 1 0 
end

I want to test whether group2, gruop3 and comb_group's AUCs are statistically significantly different from group1 (the reference group). Thus, I want 3 p-values.

I tried using the following:

roccomp truth group1 group2 group3 comb_group

But this only gives one p-value, as H0 = ll the AUCs are the same. Thus, I was thinking using the following approach instead, testing them one-by-one:

roccomp truth group1 group2

roccomp truth group1 group3

roccomp truth group1 comb_group

This may not be the most elegant way, but it is at least statistically correct?


Thank you for your time.