Indirect effects in a twolevel model PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Ulrich Schiefele posted on Tuesday, July 05, 2016 - 3:04 am
My question refers to the calculation of standardized indirect effects in a twolevel model. This is my model:

USEVARIABLES ARE
vpnt sexs sext alters teduc sinrm sinma
smast ssub;

WITHIN = sexs alters;
BETWEEN = sext teduc;
CLUSTER = vpnt;
MISSING = ALL (999);

ANALYSIS: TYPE = Twolevel;

MODEL:
%WITHIN%
sinrm ON sexs alters;
sinma ON sexs alters;
ssub ON sexs alters sinrm sinma;
smast ON sexs alters sinrm sinma;
sinrm WITH sinma;

%BETWEEN%
sinrm ON sext (a);
sinrm ON teduc (aa);
smast ON sinrm (b);
ssub ON sinrm (bb);
sinma ON sext (d);
sinma ON teduc (dd);
smast ON sinma (e);
ssub ON sinma (ee);
sinrm WITH sinma;

MODEL CONSTRAINT:
NEW (indc indcc indccc indcccc indf indff indfff indffff);
indc = a*bb;
indcc = a*b;
indccc = aa*bb;
indcccc = aa*b;
indf = d*ee;
indff = d*e;
indfff = dd*ee;
indffff = dd*e;

OUTPUT: standardized cinterval;

As a result of this input, I only get the unstandardized indirect effects. I would like to know if I can also get the standardized estimates and their corresponding p-values? Thanks.
 Bengt O. Muthen posted on Tuesday, July 05, 2016 - 6:05 pm
Try Model Indirect.
 Ulrich Schiefele posted on Wednesday, July 06, 2016 - 6:09 am
Thank you! That worked!

I only came up with MODEL CONSTRAINT because MODEL INDIRECT did not work for a similar file (see below). I don't receive any warnings and the output file does not pop up. It is there, however, and contains only the input instructions, not more.
This is part of the input file where it is still not working:

WITHIN = sexs alters;
BETWEEN = sext teduc;
CLUSTER = vpnt;
MISSING = ALL (999);
ANALYSIS: TYPE = Twolevel;
MODEL:
%WITHIN%
sinma ON sexs alters;
ssub ON sexs alters sinma;
smast ON sexs alters sinma;
%BETWEEN%
sinma ON sext teduc;
smast ON sinma;
ssub ON sinma;

MODEL INDIRECT:
smast IND sinma teduc;
smast IND sinma sext;
ssub IND sinma sext;

OUTPUT: standardized cinterval;

thanks again!
Ulrich
 Bengt O. Muthen posted on Wednesday, July 06, 2016 - 8:03 pm
There is a glitch that you avoid by changing

MODEL INDIRECT:
smast IND sinma teduc;
smast IND sinma sext;
ssub IND sinma sext;

to

MODEL INDIRECT:
smast IND teduc;
smast IND sext;
ssub IND sext;
 Ulrich Schiefele posted on Thursday, July 07, 2016 - 12:21 am
perfect, thanks, that was it!
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: