Forums  > General  > Hypothesis test of union of dependent hypothesis  
     
Page 1 of 1
Display using:  

CMPT


Total Posts: 46
Joined: May 2006
 
Posted: 2019-12-17 16:33
I have dependent correlation coefficients , and (with equal sample sizes for stationary factor returns ).

I want to test whether is neither equal to nor to .

In other words, how can I test the "combined" null hypothesis , i.e. where the alternate hypothesis is ?

PS I could test the individual hypotheses and using the T-statistic on the raw coefficients (or their Fisher-z transform) with estimated via randomizing or bootstrapping.

PPS I am interested in association in the joint left tail of and and (defined as where ) so is Spearman's rank correlation in the joint left tail.
Previous Thread :: Next Thread 
Page 1 of 1