[R] Random numbers from noncentral t-distribution

[Thomas Lumley]

On Sat, 1 Jul 2006, Long Qu wrote:

> Thank you very much for your kind reply. It solved the problem of rt( ). :D
>
>  But it seems that the qt( ) also have problems:

Yes, there does seem to be a problem near zero. A clearer version is
   curve(qt(x,df=20,ncp=3),from=0,to=0.004)
   curve(qt(10^x,df=20,ncp=3),from=-10,to=-2,n=1000)

The fix is less obvious here. I'll file it as a bug.

 	-thomas

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle