Unexpectedly, the theoretically best reject-domain of T-test is bounded.
For NHST vs , theoretically, is s.t. , rather than zero. Nevertheless, pratically a large t, rejecting both and , should not be counted as any evidence to retain or reject .
To verify the shape of –
x<-.01*(-1000:1000);
plot(x,y=dt(x-1,df=5,ncp=0)/dt(x,df=5,ncp=0),type=”l”);
###compare to the noncetrality case–
## plot(x,y=dt(x,df=5,ncp=1)/dt(x,df=5,ncp=0),type=”l”);