Re: st: sample size estimation

[saqlain raza <bhatti_sb@yahoo.com>]

Hello,
Try to get effect size by using the z transformation of correlation coefficient. But in this case, you will have to assume some value for rho that can be expected.


Saqlain RAZA
PhD Student


________________________________
From: Nick Cox <njcoxstata@gmail.com>
To: statalist@hsphsun2.harvard.edu 
Sent: Tuesday, May 15, 2012 12:20 PM
Subject: Re: st: sample size estimation

Without more detail this is an open question.

Let's trim it right down and suppose that you are interested in
summarizing treatments by their mean effects. Then there are three
pairwise differences between means, say A - B, B - C, A - C. You could
work out how big a sample you need to establish the smallest of these
three differences reliably. Then multiply up.  At a minimum you will
need to say something about how responses are distributed.

Naturally, you didn't say that you were working with means, and a
different specification would give a different answer.

Either way, the old-fashioned answer to these questions is to suppose
that that there is a cookbook formula you can plug into. A more modern
approach is to simulate the effects of conditions that are plausible
for your set-up.

Nick

On Tue, May 15, 2012 at 11:08 AM, ofran almossawi
<ofran.almossawi@gmail.com> wrote:
> Hello
> Would anyone know how to do a sample size estimation to compare the
> effect of a treatment on three groups. I can only find information on
> how to compare two groups.
>
> Regards
> Ofran
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*  http://www.stata.com/help.cgi?search
*  http://www.stata.com/support/statalist/faq
*  http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/