Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: HLM using gllamm

Subject   Re: st: HLM using gllamm
Date   Thu, 05 Dec 2002 10:17:27 +0000

Hi Derek Darves,

the level 1 variance should be given immediately
under the table for the regression coefficients under
the heading

Variance at level 1

before the variances of the random effects.

If you want a random intercept and a random slope for a
covariate x, you can use the commands

gen cons=1
eq inter: cons
eq slope: x

gllamm y x, i(id) eqs(inter slope) ...

You may find the Stata textbook examples at the UCLA Academic
Technology Services webpage useful:


These can also be accessed via the gllamm webpage -
see 'worked examples':

Best wishes,


I am trying to replicate some models in run SAS with proc mixed using
the gllamm command in Stata. At the moment, I cannot get gllamm to
print the level 1 and level 2 random effects estimates, i.e. what SAS
refers to as the "Covariance Parameter Estimates" in its output. Thus
far, gllamm reports the following output for the random effects:

***level 2 (site9)

     var(1): .598095 (.06353876)

, which I presume is for the level 2, intercept variance. How can I
also get an estimate for the level 1 residual variance, which I would
like in order to compare the amount of variance to be explained at each
level (in the intercept-only model).

Also, can someone please explain how the eq command operates in gllamm?
In corresponding with the author of gllamm, I have learned that in
order to allow the intercept to vary I must first create a constant and
link it using the eq function. How do you accomplish the same when, for
example, you'd like one of the slopes to vary randomly?


Derek Darves

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index