Mplus Discussion >> Multilevel random intercept & random slope factor

Topics
Last Day
Last 3 Days
Last Week
Tree View

Edit Profile


Multilevel random intercept & random ...

Mplus Discussion > Multilevel Data/Complex Sample >

Message/Author

psychresearcher posted on Saturday, April 23, 2016 - 2:43 am

Is it possible to have a multilevel factor model with random intercepts and random slopes for factor loadings?

This would be similar to User Manual v7 Example 9.10 (p.283) but with y1-y4 also having random intercepts and random slopes together with y5 but without the use of a covariate w?

If so, how might such a model be configured in Mplus?

Linda K. Muthen posted on Saturday, April 23, 2016 - 5:57 am

Yes. You would need a bar statement for each of the other factor indicators and to remove w from the analysis. This could be computationally heavy.

psychresearcher posted on Saturday, April 23, 2016 - 5:26 pm

Thank you Linda, would it look like this?

MODEL:
%WITHIN%

fw BY y1*-y5;

s1 | y1 ON fw;
s2 | y2 ON fw;
s3 | y3 ON fw;
s4 | y4 ON fw;
s5 | y5 ON fw;

fw@1;

%BETWEEN%

fb BY y1*-y5;

y1-y5@0;

y1 s1 ON fb;
y2 s2 ON fb;
y3 s3 ON fb;
y4 s4 ON fb;
y5 s5 ON fb;

fb@1;

Bengt O. Muthen posted on Sunday, April 24, 2016 - 10:49 am

Yes.