Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: -MVNP- Coefficient Estimates. How to standardize coefficients?


From   Roy Wada <roywada@hotmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: -MVNP- Coefficient Estimates. How to standardize coefficients?
Date   Fri, 12 Jun 2009 17:17:49 -0700

 
I'm not even sure -mvnp- produces a regression, or where you 
got it from.
 
Kit Baum's -betacoef- will only work for ols or ivreg. I think
there is a reason for that.
 
This is by hand for ols:
 
sysuse auto, clear
reg rep78 headroom trunk weight, beta
 
qui sum rep78 if e(sample)
local betaSD `r(sd)'
 
foreach var in headroom trunk weight {
     qui sum `var' if e(sample)==1
     di "`var'" _col(15) r(sd)/`betaSD' * _b[`var']
}
 
The same thing is calculated by outreg or outreg2.
 
Roy
 
 
> Thank you. But my question is about how to get the standardized
> coefficients from the non-standardized ones. I am hoping not to have
> to redo the regression itself.
>
> R Maltev
>
> On Thu, Jun 11, 2009 at 4:05 PM, R Maltev wrote:
>> To those who wrote this program or have used it:
>>
>> My question concerns the scaling of coefficients. I have used the
>> program on non-standardized explanatory variables and need to convert
>> the resulting coeffients to standardized coefficients, that is, the
>> coefficients that would have resulted if all variables had been
>> standardized (ie. sd=1). How would you suggest doing this conversion?
>>
>> Best,
>>
>> R Maltev
_________________________________________________________________
Windows Live™ SkyDrive™: Get 25 GB of free online storage.
http://windowslive.com/online/skydrive?ocid=TXT_TAGLM_WL_SD_25GB_062009
*
*   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/



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