References for type = THREELEVEL PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Peter Halpin posted on Friday, January 30, 2015 - 11:47 am
Hi,

can you please point me toward a paper describing the computational framework Mplus uses for including latent variables with type = THREELEVEL.

Thank you, Peter
 Tihomir Asparouhov posted on Friday, January 30, 2015 - 2:18 pm
We don't have a paper but it is based on the EM-algorithm.

You can see how this is done for two-level models here

http://statistics.ucla.edu/system/resources/BAhbBlsHOgZmSSIBljIwMTIvMDUvMjIvMTVfMThfMzJfNjg5X0FfTmV3X0VNX0FsZ29yaXRobV9mb3JfTWF4aW11bV9MaWtlbGlob29kX0VzdGltYXRpb25faW5fVHdvX2xldmVsX1N0cnVjdHVyYWxfRXF1YXRpb25fTW9kZWxzX3dpdGhfQXJiaXRyYXJ5X1NhbXBsZV9EZXNpZ25zLnBkZgY6BkVU/A%20New%20EM%20Algorithm%20for%20Maximum%20Likelihood%20Estimation%20in%20Two-level%20Structural%20Equation%20Models%20with%20Arbitrary%20Sample%20Designs.pdf
 Bengt O. Muthen posted on Friday, January 30, 2015 - 3:01 pm
The 3-level modeling features are described in the V7Part2 handout from our August 2012 Utrecht Mplus short course on our website.
 Peter Halpin posted on Saturday, January 31, 2015 - 3:25 pm
Great, Thank you!
 Eunsoo Lee posted on Thursday, November 10, 2016 - 8:35 pm
Hi,

I want to do Monte carlo simulation of 3-level growth model using type = THREELEVEL.

There are model command in below.

Is it possible to generate data set for 3-level growth model?

Thank you!



VARIABLE: NAMES = y1-y4 w z level2 level3;
CLUSTER = level3 level2;
WITHIN = ;
BETWEEN = (level2) w (level3) z;
ANALYSIS: TYPE = THREELEVEL;
MODEL: %WITHIN%
iw sw | y1@0 y2@1 y3@2 y4@3;
%BETWEEN level2%
ib2 sb2 | y1@0 y2@1 y3@2 y4@3;
ib2 sb2 ON w;
%BETWEEN level3%
ib3 sb3 | y1@0 y2@1 y3@2 y4@3;
ib3 sb3 ON z;
y1-y4@0;

OUTPUT: TECH1 TECH8;
 Bengt O. Muthen posted on Friday, November 11, 2016 - 10:17 am
See the UG Monte Carlo version of ex9.23 on our website (see Mplus User's Guide/Examples).
 Eunsoo Lee posted on Friday, November 11, 2016 - 8:32 pm
Thanks a lot!
I found the monte carlo version of ex9.23 and it was helpful.
However, I am trying to set the level-3 ICC for y1-y4 to a value around .05.
Can I manipulate ICC in mplus?
Thank you!

MONTECARLO: NAMES ARE y1-y4 w z;
NOBSERVATIONS = 1200;
NREP = 1;
SEED = 58459;
NCSIZES = 1[1];
CSIZES = 30[10(4)];
WITHIN = ;
BETWEEN = (level2)w (level3) z;
ANALYSIS: TYPE = THREELEVEL;
MODEL POPULATION:
%WITHIN%
iw sw | y1@0 y2@1 y3@2 y4@3;
iw*1; sw*.2; iw with sw*0;
y1-y4*.5;
%BETWEEN level2%
w@1;
ib2 sb2 | y1@0 y2@1 y3@2 y4@3;
ib2 ON w*.4; sb2 ON w*.2;
ib2*.5 sb2*.2;
ib2 with sb2*0;
y1-y4*.45;
%BETWEEN level3%
z@1;
ib3 sb3 | y1@0 y2@1 y3@2 y4@3;
ib3 ON z*.5; sb3 ON z*.3;
y1-y4*.05;
ib3*.5 sb3*.3]; [y1-y4@0];
 Bengt O. Muthen posted on Sunday, November 13, 2016 - 6:32 pm
icc's are most easily manipulated by changing the between-level variances.
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: