I have a path analysis for two different study groups (stgrp) [A , B]. Should I start with study group incorporated into the model or as multiple group analysis. I am not including the factor loadings in the example for keeping it simple.
Incorporated into model: posemo on stgrp; negemo on stgrp; ident on stgrp; engage on posemo; engage on negemo; engage on ident;
as multiple group: grouping is stgrp (1=interv 2=control); engage on posemo; engage on negemo; engage on ident;
I would do each group separately to make certain the model fits in each group. Then I would test for measurement invariance of the latent variables. After establishing measurement invariance, I would test for equality of structural parameters. This process is described in the Topic 1 course handout and video.
Ossi Pesämaa posted on Thursday, November 11, 2010 - 8:47 am
I am new with Mplus. I have four latent factors and thirteen variables. I try to test if the latent measures are equivalent across two waves and if one of them mediate the effects on a third latent measure. I tried to find a syntax for testing equivalence (free, factor loadings equal, error, and fixing structural paths one by one) but did not find it. Can you give me any advice here? The syntax input look something like this.
--- DATA: FILE IS test.dat;
VARIABLE: NAMES ARE ID x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x11 x12 x13 group;