Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

st: Interpreting random effect variance using GLLAMM versus GLLAPRED


From   "Denise Duffy" <[email protected]>
To   <[email protected]>
Subject   st: Interpreting random effect variance using GLLAMM versus GLLAPRED
Date   Sun, 2 Jan 2011 11:41:44 -0500

I'm using GLLAMM to estimate an item response type model with individual random effects and am having trouble interpreting the random effect variance reported by GLLAMM.  After the model converges, the GLLAMM results report a variance of the random effect :

Variances and covariances of random effects
------------------------------------------------------------------------------
***level 2 (newid)
 
    var(1): .33895169 (.24571358)


I then estimated the individual random effects using gllapred (gllapred u, u).  The variance of the mean estimate does not equal the GLLAMM-reported variance (.33895169).  I would expect these to be equivalent.  I find that this occurs whether or not the GLLAMM-reported variance is statistically significant.  If anyone has insights into how to interpret these random effect variance calculations (why they would be quite different), I'd appreciate your feedback.

Best regards,
Denise Duffy
Institute for Social Research
The University of Michigan



*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index