Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Re: Re: ST: one-tailed test


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Re: Re: ST: one-tailed test
Date   Mon, 11 Oct 2004 14:34:29 +0100

Not precise??? 

In cases in which the P-value is left behind as 
a returned value (see results of -return list- or 
-ereturn li-), dividing by 2 will give you 
several decimal digits of precision. 

As usual, that precision could be spurious if underlying 
assumptions are not valid, or if data are coarsely or poorly 
measured, etc., but you can get a lot of precision 
out of Stata. 

In cases in which there is no such returned P-value, 
you may have to dig down and fire up a function 
or two, but the precision is there in principle. 

The key point is that you are _not_ limited to 
taking what is printed out and dividing _that_ 
by two. 

Nick 
n.j.cox@durham.ac.uk 

Levy Lee
 
> thanks again, 
> 
> but i do not think this is precise, are there other way to go?
> 

Eric Uslaner
 
> > Levy Lee asks how to perform a one-tailed test after a 
> regression and
> > Richard Williams responds: "ust look at the printout, and 
> see (a) is 
> > the coefficient for x2 positive, and (b) is it significant 
> at the .10
> > level  (or double whatever alpha level it is you want to 
> chose).  You
> > don't need 
> > to do any additional analyses, the information you need is already
> > there."
> > 
> > Well, not quite.  Stata (annoyingly) always reports 2-tailed tests,
> > whereas one-tailed tests are clearly more appropriate.  So 
> you need to
> > divide the p level by 2.

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index