Statalist


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

Re: st: you go -xtmelogit-


From   George Hoffman <ghoffman@mcw.edu>
To   <statalist@hsphsun2.harvard.edu>
Subject   Re: st: you go -xtmelogit-
Date   Thu, 25 Dec 2008 07:44:56 -0600

Jay -
Of course not the data. I'm one of those non-methodologists trying to use
sophisticated tools on complex data, learning as I go - so I was hoping to
learn from your model specification- not to detract from the expertise of
the professionals.
Thanks!
GH


> From: jverkuilen <jverkuilen@gc.cuny.edu>
> Reply-To: <statalist@hsphsun2.harvard.edu>
> Date: Thu, 25 Dec 2008 02:10:37 -0500
> To: <statalist@hsphsun2.harvard.edu>
> Subject: RE: st: you go -xtmelogit-
> 
> I can't give out the data (confidentiality) but I'll put up the model when I
> get the chance. The model's nothing special, just two random effects and some
> fixed effects predictors. The example in the manual is probably better
> 
> The main thing is the fact that what is known to be a computationally tricky
> model works well enough without "fine tuning" to ponder turning its routine
> estimation over to non-methodologists.
> 
> 
> 
> -----Original Message-----
> From: "George Hoffman" <ghoffman@mcw.edu>
> To: statalist@hsphsun2.harvard.edu
> Sent: 12/24/2008 5:37 PM
> Subject: Re: st: you go -xtmelogit-
> 
> Jay - it would be interesting to me (and maybe to others?) to see the model
> command you used. I'm sure most on this list have great experiences with
> stata - thanks for taking the time to write it donw.
> 
> GH
> 
> 
> 
>> From: "Verkuilen, Jay" <JVerkuilen@gc.cuny.edu>
>> Reply-To: <statalist@hsphsun2.harvard.edu>
>> Date: Mon, 22 Dec 2008 16:17:04 -0500
>> To: <statalist@hsphsun2.harvard.edu>
>> Subject: st: you go -xtmelogit-
>> 
>> I just figured StataCorp would appreciate this:
>> 
>> I ran a crossed random effect binary model today on some data I got this
>> afternoon, which has 72 observations per subject, 120 subjects, 3
>> conditions for the study variable that's random.
>> 
>> It ran on the first try, in under 15 minutes.
>> 
>> Now obviously this involves having "good" data with pretty big effects,
>> but as anyone who ran these kinds of models even a few years ago can
>> attest, estimating a crossed random effect is no easy task, but
>> -xtmelogit- did everything with no sweat, no messy model specification,
>> etc. 
>> 
>> Thanks, and Merry Xmas a few days early (or is it Happy Hanukah on
>> time).
>> 
>> JV
>> 
>> 
>> 
>> *
>> *   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/



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