|
|
 |
| Freeing latent means within multiple ... |
 
|
| Message/Author |
|
|
|
|
Hi,
I am currently testing for measurement invariance over time and across groups as part of a multiple group autoregressive panel model analysis.
However, when I try to test scalar invariance over time, the latent means remain constrained at zero at all time points within the first group despite me needing to free them in the test. Below is an illustrative example of my code in case that helps make things clearer where Y1 (X1) and Y2 (X2) are the same construct at different time points:
Y1 by y11 y21 y31 (L1-L3);
Y2 by y12 y22 y32 (L1-L3);
Y1@1 Y2@*;
[Y1@0 Y2*];
X1 by x11 x21 x31 (L4-L6);
X2 by x12 x22 x32 (L4-L6);
X1@1 X2*;
[X1@0 X2*]; !The mean of X2 remains constrained at zero despite being freed
[y11 y21 y31] (I1-I3);
[y12 y22 y32] (I1-I3);
[x11 x21 x31] (I4-I6);
[x12 x22 x32] (I4-I6);
model group-2:
...
Is there any way of overriding the default for MPlus to constrain the means to zero in the first group, as just using * doesn't seem to work? |
|
|
|
|
| Use Model-group1 to set the factor means the way you want them. |
|
|
|
|
| Okay great, got it working. Thank you! |
|
|
| Back to top |
|
|
