Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: SAS to Stata


From   "Stas Kolenikov" <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: SAS to Stata
Date   Fri, 22 Aug 2008 09:48:58 -0500

On Fri, Aug 22, 2008 at 9:42 AM, Raphael Fraser
<raphael.fraser@gmail.com> wrote:
> I think I found the solution but while the covariance for the random
> effects parameters are the same the variances are not. I can't imagine
> that they are using different methods! BUT some thing is off here. The
> coefficients and standard error for the fixed effects are exactly the
> same. Is there a reasonable explanation?

Different approximations are a very plausible one. The fixed effects
are much easier to estimate, and you can get some very reasonable
answers even with -regress, cluster-. With random effects, you might
have some weird parameterizations, where covariances are estimated
separately from variances, and may happen to have different
multipliers like 1/N or 1/(N-1) -- I am just thinking aloud here.
Bobby G. might provide a better explanation, if there is one :)).

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
*   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