SEMLj cluster option
Posted: Thu Sep 02, 2021 1:41 pm
Hi,
I am amazed about the new package SEMLj allowing to estimate SEMs. However, similar to JASP, there is one important option I am missing: If you have longitudinal data (=LONG format), but are not explicitely interested in longitudinal effects, you can specify a cluster variable for the subjects in LAVAAN and will then get robust Hubert-White-Standard errors, and hence you are then able to estimate correct SEMs for longitudinal data. I assume this should not be hard to implement in SEMLj - and would be very happy to see it in an upcoming version.
Cheers and thanks for all your good work,
Steffen
I am amazed about the new package SEMLj allowing to estimate SEMs. However, similar to JASP, there is one important option I am missing: If you have longitudinal data (=LONG format), but are not explicitely interested in longitudinal effects, you can specify a cluster variable for the subjects in LAVAAN and will then get robust Hubert-White-Standard errors, and hence you are then able to estimate correct SEMs for longitudinal data. I assume this should not be hard to implement in SEMLj - and would be very happy to see it in an upcoming version.
Cheers and thanks for all your good work,
Steffen