Hi all, I appreciate your help, I attach a model made in SPSS Amos, I think this is graphically correct, but I get the error described in the question, I appreciate your comments !!
Dear Caleb Terrel Orellana, the way you are trying to analyze your constructs is wrong. I have a question: are you trying to test scale or relationships?
Adil
Dear Adil Bilal, Thank you, why do you think it's wrong? , And I'm testing relationships
1) Try to erase the regression weights 1 from: PSQIQ to PSQ; PSQOC to PSQ; PSQEQ to PSQ
2) Where are the observed variables of your PSQ latent variable? It seems that they are missing.
Hello Lisete thanks:
1. I already deleted the weights you mentioned and the error in Amos continues
2. PSQ is a formative variable of third order
Another question I have is: that I get results but when in "Analysis Properties" I check "Try to fit unidentified models" ... attached
ok, is formative, but I only can see he latne construct (where are the observed variablkes for PSQ?) PSQ is measured only by 1 variable?
The model you're fitting is a second order factor model with second order latent variables psqiq, psqoq, and psqeq . That part of your model is identified.
Then, you want to create a latent variable model with construct psq and psqiq, psqoq, and psqeq as formative indicators of psq. That's the part of your model that is not identified. I see wome covariances between psqiq, psqoq, and psqeq being fixed to one. Free them. Then, I believe you need a reflexive indicator of psq for the model to be identified. In any case, all you need to do is check how to identify a model with formative indicators, because that's your problem.
I thought we can't do formative measurement models with CB-SEM (such as AMOS)?
I've read some comments in other threads that PLS (such as Smart PLS) would do better for formative indicators.
Hi Caleb Terrel Orellana ,
Can you please explain me how you got the results? I have almost same type of model and experiencing the same problem.
I am also have the same error because each latent factor should have one indicator path with a regression constraint to 1. Then I delete the path that have the error and draw it again. Then it can solve my problem.
I hope this web site can help you to solve your problem. https://ruvideos.org/efC81f-Z22Q-видео.html
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size