Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Mean Test


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Mean Test
Date   Sun, 22 Jan 2006 22:51:05 +0000 (GMT)

Rijo wrote:
> Isn't command
>
> foreach var of varlist popgrowth-safewater {
>         oneway `var' region, t sid
>        }
> doing multiple comparison tests for one variable at a time?
> thats exactly what I want to do. Coz what I am doing is
> comparing means of say 30 different variables for 4 groups. I
> mean at any given time I will be comparing the differences in
> means of 4 groups for A particular variable. And I want to get
> the p-values for all possible combinatins. I mean, not just the
> p value of the H_0: mean1=mean2=mean3=mean4
>
> But also the p-values for other H_0 such as mean1=mean2, 
> mean2=mean3
> and so on. (total 6 comparisons.)

Sorry, I am going to give you an answer you are probably not going to like: I would seriously
advise you against doing 6*30=180 test. If you want to explore your data, than graphs are by far
more advisable than tests. Tests are good for testing your theories, graphs are good for exploring
your data. 

> A slightly different question:
> is there a way one can make stata to report p values upto say 5
> or 6 decimals in these kinds of tests or report it in a 
> different format? I have a very large data set and my p-values
> many times turns out to be 0.000. It looks odd to report a
> value like 0.000 in a table.

The p-value is the type I error (finding a significant result when you shouldn't) rate. This type
I error rate is distorted when you do many tests without theory. Sidak's correction is an attempt
to correct that, but it is at best an approximation. There are other approximations, like the
Holm's and Bonferroni's correction. So the p-values should not be taken as exact to the third of
fouth or sixteenth decimal place. Anyhow if it is less than .05 (or .01 if you like) than it is
significant, so more decimals will give you no extra information.

HTH,
Maarten

-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting adress:
Buitenveldertselaan 3 (Metropolitan), room Z214

+31 20 5986715

http://home.fsw.vu.nl/m.buis/
-----------------------------------------



		
___________________________________________________________ 
To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com
*
*   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 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index