RE: st: Two way table of regression coefficients or intercepts with -esttab-

["Jesper Lindhardsen" <JESLIN01@geh.regionh.dk>]

Hi Richard,

I would use -estout- with the append option, and regress (or other commands) in the relevant subgroups through loop.
Building on your own example, something like this.....

eststo clear

levelsof a_q,local(l_a_q)
levelsof b_q,local(l_b_q)

capture erase yo.txt

foreach l_a of local l_a_q {

	est clear
	
	foreach l_b of local l_b_q {
		eststo e_`l_a'_`l_b': reg c if a_q==`l_a' & b_q==`l_b'
	} 

	estout * using yo.txt, append collabel(none)

}

If the command you want to use do not return estimate(eg summarize), then you can use -estpost- to make up for for this...
In my experience, it not worth the effort to rely on any output command to give you a perfect final table. IMHO, It's way easier to get the results in the right rows and columns, and then use excel or other spreadsheets as an intermediary before a wordprocessor.

HTH, Jesper

Jesper Lindhardsen
MD, PhD candidate
Department of Cardiovascular Research
Copenhagen University Hospital, Gentofte
Denmark







-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Richard Herron
Sent: 23 September 2011 16:28
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: Two way table of regression coefficients or intercepts with -esttab-

-esttab- and -eststo- are from the -estout- package.

http://repec.org/bocode/e/estout/index.html

Thanks!

On Fri, Sep 23, 2011 at 10:07, Nick Cox <njcoxstata@gmail.com> wrote:
> Please remember the request to explain where user-written commands you
> refer to come from.
>
> Nick
>
> On Fri, Sep 23, 2011 at 2:58 PM, Richard Herron
> <richard.c.herron@gmail.com> wrote:
>> Can I make a two-way table of regression coefficients with -esttab-?
>> Sometimes I like to make a two-way table with either intercepts or
>> coefficients and I would like to automate it.
>>
>> For example, here I generate three random variables, then form
>> quintiles on -a- and -b-, then find the mean of -c- in each of the 25
>> groups (intersections of -a_q- and -b_q-) with -regress-. This is
>> fairly easy with -table-, but is less extensible to displaying
>> intercepts, significance, etc (and I am not sure how to generate a
>> LaTeX table from -table-). Can I so this with -esttab- or -estout-?
>>
>> Thanks!
>>
>> * --- begin code ---
>>
>> * generate data
>> clear
>> set obs 2000
>> generate a = rnormal()
>> generate b = rnormal()
>> generate c = rnormal()
>>
>> * generate quantiles for for a and b
>> xtile a_q = a, nquantiles(5)
>> xtile b_q = b, nquantiles(5)
>>
>> * I would like something like this two-way table of means, but that is
>> extensible to intercepts and/or more coefficients
>> table a_q b_q, c(mean c)
>>
>> * but the best I can do is this very wide table that is not very readable
>> eststo clear
>> bysort a_q b_q: eststo: quietly regress c
>> esttab, not noconstant nogap compress
>>
>> * --- end code ---
> *
> *   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/