Your model is fairly large given for the rather small sample size. This could be part of the problem. Also, and perhaps more importantly, you have many indicators for each factor. It seems unlikely that these indicators are unidimensional within each factor. In other words, your model may be misspecified, which could lead to the problems that you encounter. I would try a simpler model with fewer indicators and/or fewer latent variables.

Do you have missing values, and if so how did you deal with them? If the matrix is not positive def it could be either as Christian Geiser says related to the small sample size, but can also happen with imputing. I assume you have saved the PHI matrix. What is its rank, does it have an inverse, etc. ?

Dear Mr. Christian Geiser

Thank you so much for your answer. I will try to get rid of some indicators and get more samples.

Dear Mr. Daniel Wright

Thank you so much for your answer. Unfortunately, I don't really know how to explain it in English and my understanding regarding this topic is still lacking. But here is the output from my model. Any help to fix this problem would be much appreciated.

Thank you so much.

Problem will be the correlations of independent variables. Three are larger then 1.00. These are the correlations of RL and E, AE and SAT. These result in the message matrix not positive definite.

Dear Mr. Peter Moorer

Thank you so much for your answer. How do I fix these correlations? I would be grateful if you could help me in any way.

Just put ADD=OFF on syntax