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

st: RE: How to select variable for multivariate analysis


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: How to select variable for multivariate analysis
Date   Wed, 8 Sep 2004 12:58:25 +0100

Just from what you tell us, with nothing 
said about what the data are, what the 
scientific problem is, or what method 
you are applying, then there has to be
a question mark against variable c 
and a slightly smaller one against variable b 
and a slightly smaller one yet against variable a. 
In fact, a conservative approach might reject them all. 

To say this is to say very little, however. 

All sorts of things could be going on, and 
p-values alone won't tell you what's  
a good analysis. It is possible that 
the variables are just fighting with each 
other and p-values will become more respectable 
if one or two variables are left out. It 
is possible that results will be changed
drastically if you look at one or more 
of the variables on a transformed scale. 
And so on. 

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

Syed O Masood
 
> I have a technical question. Do we have to use p-value
> for selecting independent variables fro multivariate
> analysis? 
> 
> Example:
> 
> I have 3 independent variables a, b, c which I have to
> regress on dependent variable y. 
> 
> In independent variable, the p-value for variable a is
> 0.09, varable b is 0.23 and variable c 0.77. If I have
> to do multivariabte analysis, is it necessary to
> reject variable c because the p-value is 0.77??
> 
> In multivariate analysis, does order of entering
> variables make a difference??

*
*   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