Dear Abderrahim,

I guess you're dealing with a "model identification" issue (which is one of the most important steps along with model "specification" and "evaluation" in your SEM analysis).

AMOS is basically telling you that you need to better "identify" your model: perhaps it's a problem of model saturation, meaning that you have no degrees of freedom left (df = 0) in your hypothesized framework. In other words, you need to add/delete parameters, otherwise your model will be equal to the "saturated" model with Chi-square = 0 and df = 0, which is an "untestable" model that usually serves as the "baseline" comparison to compare fitting indexes (e.g., relative fitting indexes as the CFI).

Hence, try to add/delete indicators and correlate your construct with other latent variables in your AMOS diagram, so that you "free" parameters and let the software compute the SEM analyses.

Have a look at this "SEM identification procedure" video: https://youtu.be/JkZGWUUjdLg

Hope this help,

best luck with your research!

Thanls a lot Dear Lamberto Zollo for your explanation