qrprocess, Bonferroni correction

[Masoumeh Sanagou]

StataList,

After doing

qrprocess A i.B, quantile(0.01(0.01)0.99) method(qreg)

to find the significant results
should the p-values be compared to 0.05?

or

should be the p-values be compared to 0.05/97=0.0005 as it estimated multiple quantile regressions simultaneously?
(Bonferroni correction, number of estimated percentile or number of comparisons=0.01(0.01)0.099=97)


Regards,

[Carlo Lazzaro]

Masoumeh:
I'm not familiar with -qrprocess- but what follows can be generalized to any multiple comparison that includes Bonferroni's correction.
You're right with stating that, if you have 97 multiple tests to perform, the arbitrary 0.05 should be divided by 97 (=0.0005).
Therefore, any comparison that reaches a p-value <0.0005 is actually significant at 0.05.