2-Level Seemingly Unrelated Regression PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Jorge Walter posted on Tuesday, July 30, 2013 - 8:38 am
I'm trying to run two 2-level regressions while accounting for correlated errors between the regressions (i.e., a seemingly unrelated regression). The two regressions have different DVs and a mix of the same and different IVs. Following a previous post on a related topic (http://www.statmodel.com/discussion/messages/11/2964.html?1205362039), I've adapted your standard 2-level model as follows:

VARIABLE: NAMES = y1 y2 x1 x2 x3 w clus;
WITHIN = x1 x2 x3;
BETWEEN = w;
CLUSTER = clus;
DEFINE: CENTER x1 x2 x3 (GRANDMEAN);
ANALYSIS: TYPE = TWOLEVEL;
MODEL:
%WITHIN%
y1 ON x1 x3;
y2 ON x2 x3;
%BETWEEN%
y1 ON w;
y2 ON w;
OUTPUT:
STANDARDIZED;

Now my question is: Does this model account for correlated standard errors between the two regressions and if yes, how?

Thanks in advance for your help!
 Linda K. Muthen posted on Tuesday, July 30, 2013 - 4:18 pm
Check if the parameters y1 WITH y2 are estimated as the default. If not add them to your model.
 Jorge Walter posted on Tuesday, July 30, 2013 - 5:47 pm
Thanks so much, Linda, for your fast reply. If I understand you correctly, then, the inclusion of "y1 WITH y2" into the %WITHIN% portion of the above model would account for correlated standard errors between the two regressions, correct?
 Linda K. Muthen posted on Tuesday, July 30, 2013 - 7:41 pm
The correlated residuals not standard errors.
 Jorge Walter posted on Tuesday, July 30, 2013 - 10:39 pm
Makes sense, thanks so much for your help!
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: