mardi 3 mars 2015

About calculating p-value

I have a problem in understanding how to calculate the p-value using a bootstrap (or numerical approach) as in this articl : http://ift.tt/1BC4MaS page 13 he say :


The inference about the test statistics is to compute P-value as the fraction of the bootstrap test statistics which is greater than the original one.


It is easy to somput the original test statistics and the vector of test statistics in the numerical approac, but I don't understand how to determin the p-value.


Thank you for help.


Aucun commentaire:

Enregistrer un commentaire