Hello!
Im having trouble with running this model and i have no idea what to do. I keep on getting the message that it is unidentified.
The problem is with the factors CL CS SQ they have correlations amongst each other but in a structural model you need to identify exogenous and endogenous variables and CL and CS are neither of them!
Hey Andrea, how you doing? Long time no see on Semnet! :))
Javier de la Fuente I would throw another hypothesis into the basket:
a) I do not know AMOS so much but is it correct to have all these "1"s as factor loadings? Are these fixations?
b) SR has only two indicators. In case your second order factor is misspecified and has a non-sing. loading/effect on SR, this part will be locally unidentified.
Best,
Holger
Dear Javier,
As Holger mentions in "1)", I also think that you need to keep only one factor loading per factor set as "1". This means that this particular loading will not be estimated but set to "1". To erase the excess of 1s, use the right click to select one factor loading arrow at a time, and in the dialogue which will open, you need to erase the "1" in the parameter box, leaving only one 1 per factor! (try to choose the item or observed measure with the strongest loading on the latent factor, as this produces output easier to interpret).
Actually, I think you are not giving the software the freedom to estimate the loadings (they're all set to =1). Leave all the 1s for errors, but leave only one factor loading = 1 per latent factor.
Not sure if the second order factor also needs one 1 on one of its three loadings (may be).. I found this book useful: Byrne, Barbara M. (2010). Structural Equation Modeling with AMOS. Basic Concepts, Applications, and Programming (2nd ed).
Hope this helps.
Best regards, Rafael
(PS1: Amos has its weirdness... the SRMR you will find under Plugins and you need to open it first, and run the analysis with that box open for the software to give you that output, which does not appear in the general output. Feel free to DM me if you struggle. Rafa.)
Hi Rafael,
yes you are right: One of the second order loadings has to be fixed to 1.
Best,
Holger
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running