Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: how to escalate coefficients using outreg2 or a similar program?


From   Benjamin Villena <b_villena@yahoo.com>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   Re: st: how to escalate coefficients using outreg2 or a similar program?
Date   Thu, 18 Oct 2012 11:33:23 -0700 (PDT)

Dear John (or anyone else who knows the answer)
The example you gave is great, but I am using a multiequation estimation command called cmp (Roodman 2012)
How can I adapt your example if I have to estimate this multiequation model with cmp using standardized regressors ?

cmp (y1= x1-x10 z) (y2 = x1-x10 z)  (y3= x1-x10 z) (y4= x1-x10 z) (y5= x1-x10 w)  if ss=1, `options'

 In other words, how can I write the 'syntax' line of my new program to be able to invoke cmp with the right syntax within the new ado file?

Thanks,


----- Original Message -----
> From: John Luke Gallup <jlgallup@pdx.edu>
> To: statalist@hsphsun2.harvard.edu
> Cc: 
> Sent: Wednesday, October 17, 2012 6:44 PM
> Subject: Re: st: how to escalate coefficients using outreg2 or a similar program?
> 
> -outreg2- (and -outreg-) take all their information about the coefficients 
> directly from the e(b) and e(V) matrices, so you _have_ to modify these matrices 
> if you want to report non-standard estimates.  These matrices can only be 
> modified by an -eclass- program, so you must create a program within your .do 
> file to do this.  In Benjamin's case, it sounds like he has already created 
> a new estimation command, so all he has to do is use -ereturn post- to save the 
> results in the e(b) and e(V) matrices before running -outreg2-.
> 
> pseudocode to do this would look like:
> 
> program escalate, eclass
>    version 12.1
>    syntax varlist [if] [in] [fw aw] [, *]
>    oldestimationcommand `varlist' `if' `in' `weigth' `exp', 
> `options'
>    tempname newb newV 
>    matrix `newb' = e(b)
>    matrix `newV' = e(V)
>       ... code to change newb and newV ...
>    ereturn post `newb' `newV' ... 
>     [-ereturn post- clears all the previously posted e() information, so 
> everything you want to keep has to be reposted in the -ereturn post- command]
> end
> 
> John
> 
> On Oct 17, 2012, at 7:45 AM, Benjamin Villena <b_villena@yahoo.com> wrote:
> 
>>  Hello Statalist members,
>> 
>>  In order to numerically stabilize an estimation procedure that uses 
> complicated numerical integration, I need to standardize my regressors X. 
>> 
>>  I plan to use outreg2 to put  several specifications in a table using the 
> same estimation method. 
>> 
>>  However, I want to report escalated coefficients and standard deviations 
> instead of the ones I get directly from my estimation routine. In other words, I 
> want to divide each coefficient and its standard error by the sample standard 
> deviation of the associated regressor after the estimation, and report them in a 
> table.
>> 
>>   I guess I could manipulate e(b) and e(V) in order to get what I want, but 
> then I think I could not use outreg2 to create a table. Is there a simple way to 
> do this using already built code such as outreg2 or similar?
>> 
>>  Thanks,
>> 
>>  *
>>  *   For searches and help try:
>>  *  http://www.stata.com/help.cgi?search
>>  *   http://www.stata.com/support/faqs/resources/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/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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