st: calculation of standard error of rho under "xtreg, mle" command

[Obi Ukoumunne <obioha.ukoumunne@kcl.ac.uk>]

Dear Statalist

My query concerns the random effects regression procedure in Stata.

I've been using the "xtreg" procedure (with the "mle" estimation option) in 
Stata to obtain maximum likelihood estimates of the intraclass correlation 
coefficient (rho). A standard error is provided for the estimate of rho as 
well as the between and within cluster variance components from which rho 
can be calculated.  I have two questions:

a)  Which formula is used to obtain the standard error of rho.  I couldn't 
find this in the Reference Manual

b)  If the standard error is likelihood based, from the Hessian matrix, I 
was wondering how it is obtained since I understood (from the Stata manual) 
that the log-likelihood function is expressed in terms of the between and 
within cluster variance components and would have thought likelihood-based 
standard errors could only be obtained for these parameters.

Many thanks in advance.

Obi

Mr Obioha C. Ukoumunne
MRC Fellow
Department of Public Health Sciences
King's College, London
5th Floor
Capital House
42 Weston Street
LONDON SE1 3QD

Tel: 020 7848 6632
Fax: 020 7848 6605

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/