Dear Scholar, I have carried out Path Analysis in AMOS for testing the model fitness for the Structural Equation Model having Formative Constructs. The model has 9 Observed Exogenous variables and 4 observed Endogenous variables, 5 Error terms, and one Dependent Variable. sample size 400. The test result says “Minimization was unsuccessful”. The results that follow are therefore incorrect. The model is probably unidentified. In order to achieve identifiability, it will probably be necessary to impose 2 additional constraints. Function of log-likelihood = -8283.237. A number of parameters = 63”.
Why AMOS output gives the result as an Unidentified model? and How to impose 2 additional constraints?
May I request the Scholars to guide me to make the model as Over identified model.
Firstly, I think AMOS can not handle formative measurement model ( I read somewhere I have not reference for it )
Next, i guess you are having high linearly issues and/or negative covariance etc.
If you want I can check your model.
Good luck
Hello Arumugam,
Identification depends on: (a) the number of observed/manifest variables in your data set; and (b) the total number of parameters to be estimated. The AMOS program feedback that you get would ordinarily suggest an under-identified situation (more parameters to be estimated than data points or observations). But, your query post indicates either 13 or 14 observed variables (not really clear), and 63 apparent parameters to be estimated. The observations (assuming 13 variables) would be: 13 * 14 /2 = 91, which is more than 63, so the model should be identified.
I'd guess that what has happened is that AMOS expects one of the variables linked to a latent variable to have a fixed loading of "1" assigned (by you), and that you haven't done this for one or more variables. Without seeing your model, I can't offer any more specific advice.
Good luck with your work.
Dear David Morse
Is it right to do analysis of formative construct by PLS not AMOS (CB-SEM) ?
Best,
Hello Alireza,
That depends on the nature of the variables, the specific research questions, and, to some extent, philosophical perspectives. The PLS-SEM believers ascribe many virtues to the method (https://www.smartpls.com/documentation/choosing-pls-sem/pls-sem-compared-with-cbsem), though each has its own place. Without knowing more about your specific aims and variables, I can't really advise you as to which way is more appropriate.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Papers