Re: st: testing coefficients using parmby command

["Roger B. Newson" <r.newson@imperial.ac.uk>]

There are a large numbers of assorted -parmest- examples in the on-line 
help for -parmest-, and in the various references cited there. However, 
in your case, the trick is probably to use -levelsof- to get a list of 
levels of -id-, and to loop over those, storing the temporary filenames 
in a global macro. The resultssets can then be concatenated, and the 
-idstr()- option can be used for storing information specific to each 
resultsset on one or more resultsset-specific quantities. The -split- 
command can be usefful, if we have multiple resultsset-specific 
variables. As in:

#delim ;
global tflist "";
levelsof id, lo(ids);
foreach ID of num `ids' {;
  regr y x1 ,robust;
  test (_cons=0) (x1=1);
  scal stestp=r(p);
  tempfile tfcur;
  parmest, label idstr(`ID'&`=stestp')
    saving(`"`tfcur'"', replace) flis(tflist);
};
clear;
append using $tflist;
split idstr, parse(&) gene(S_);
destring S_1, gene(id);
destring S_2, gene(testp);
drop idstr S_*;
sort id, stable;

This will create an output dataset (or resultsset) in the memory, with 1 
observation per fitted parameter and 2 extra variables -id- and -testp-, 
containing the appropriate value of -id- and the test P-value, respectively.

Many variations on this theme are possible. For instance, if -id- has 
value labels and/or a variable label orand/or a format, then we might 
use the SSC package -descsave-, with the -dofile()- option, to re-create 
these in the concatenated resultsset.

I hope this helps.

Best wishes

Roger


Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
UNITED KINGDOM
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Email: r.newson@imperial.ac.uk
Web page: http://www.imperial.ac.uk/nhli/r.newson/
Departmental Web page:
http://www1.imperial.ac.uk/medicine/about/divisions/nhli/respiration/popgenetics/reph/

Opinions expressed are those of the author, not of the institution.

On 27/07/2012 13:14, David Ashcraft wrote:
> Thanks Roger. Do you have any example that I can follow? What you have said makes complete sense but it is a little overwhelming for me to implement.
> Regards
>
> David
>
>
> ----- Original Message -----
> From: Roger B. Newson <r.newson@imperial.ac.uk>
> To: statalist@hsphsun2.harvard.edu
> Cc:
> Sent: Friday, July 27, 2012 1:22:38 PM
> Subject: Re: st: testing coefficients using  parmby command
>
> To do this, you have to use -parmest- inside a loop, instead of using
> -parmby-. You need to loop over the values of -id-. In each iteration of
> the loop, you nee to execute your command for the current -id- group,
> use -test- to extract the test statistics for the current -id- group,
> and then use -parmest- for each -id- group, storing the test statistics
> in the -idnum()- or -idstr)_- options of -parmest-, and storing the
> results for the current -id- group in a -tempfile-. After doung this,
> you can concatenate the resultssets created, using -append- or maybe
> -dsconcat-, and then you will have a combined resultsset in memory,
> complete with test statistics.
>
> I hope this helps.
>
> Best wishes
>
> Roger
>
>
> Roger B Newson BSc MSc DPhil
> Lecturer in Medical Statistics
> Respiratory Epidemiology and Public Health Group
> National Heart and Lung Institute
> Imperial College London
> Royal Brompton Campus
> Room 33, Emmanuel Kaye Building
> 1B Manresa Road
> London SW3 6LR
> UNITED KINGDOM
> Tel: +44 (0)20 7352 8121 ext 3381
> Fax: +44 (0)20 7351 8322
> Email: r.newson@imperial.ac.uk
> Web page: http://www.imperial.ac.uk/nhli/r.newson/
> Departmental Web page:
> http://www1.imperial.ac.uk/medicine/about/divisions/nhli/respiration/popgenetics/reph/
>
> Opinions expressed are those of the author, not of the institution.
>
> On 26/07/2012 12:34, David Ashcraft wrote:
> > I am using parmby command to regress a number of equations by id as per below:
> >
> > parmby "regr y x1 ,robust",by(id) label saving(myparms3,replace)
> >
> > Now I want to test for each id:
> >
> >
> > test _cons=0
> > test X1=1
> > test (_cons=0)(X1=1)
> >
> >
> > and store these results on myparms3 file.
> >
> > At the moment when I run the last three commands, it simply give me the result of last regression.
> > Any help will be highly appreciated.
> > Regards
> >
> >
> > David
> >
> > *
> > *   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/
*
*   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/