It is very difficult to say what's going on without seeing your model. Can you post your Mplus MODEL statement?

Thank you professor for your kindly response. My model statement is:

TITLE: SEM;

DATA: FILE NAME IS C:\Users\Saeid\Documents\Mplus\SAEID.DAT;

VARIABLE: NAMES ARE JS1 JS2 JS3 JS4 JS5 JS6 JS7 JS8 LE1 LE2 LE3 LE4 talash; USEVARIABLES ARE JS1-JS8 LE1-LE4; CLUSTER = TALASH;

ANALYSIS: TYPE = TWOLEVEL; H1ITERATIONS = 10000; ESTIMATOR = MLF;

MODEL: %WITHIN%

JSW BY JS1-JS8; LEW BY LE1-LE4; JSW ON LEW;

%BETWEEN%

JSB BY JS1-JS8; LEB BY LE1-LE4; JSB ON LEB;

OUTPUT: Standardized;

If you want I can send you my data-set.

This is quite a complex multilevel SEM model with many indicators. It is difficult to know why it doesn't work. I would start simpler. For example, look at just one factor at a time and examine whether the factor model works at both levels for each factor separately.

Note that the Level-2 factor structure need not be the same as the Level-1 factor structure.

Also, how large is your Level-2 sample size? In multilevel SEM, problems often occur due to the Level-2 sample size being too small. As a rule of thumb, there should be > 100 level-2 units (clusters).

Christian Geiser by your explanation I can write down my statement like this:

......

model: %within%

JSW BY JS1@0 JS2@0 JS3@0 JS4@0 JE5@0 .....JS8@0

LEW BY [email protected]@0;

JSW ON LEW;

%BETWEEN%

JSB BY JS1-JS8; LEB BY LE1-LE4; JSB ON LEB;

and then reverse the model. Am I right?

my sample size is about 250.

You might want to have a look at my answer to this question:

https://www.researchgate.net/post/Iterations_stopped_due_to_small_change_in_likelihood_function_in_hlm

Dear professor Kelvyn Jones

Thank you so much for your kindly response.