Hi,

it is not hidden; it simply has no t-value. It is fixed -- thus not estimated. No estimation, no uncertainty, no null hypothesis test, no t-value.

What I wonder about: How did you identify the independent latent variable? Its first loading should be fixed, too.

Beyond that, I would rather be concerned about the misfit of the model. The huge difference in t-values (which should reflect strong differences between the factor loadings) could signal that the structure of the independent variable is more diverse--but that depend on the question wordings. So, enough teaching ;)

Best,

Holger

Holger Steinmetz if you run each model in LISREL the result is same. the picture i shared, was only a sample. I'm sure you did not work with LISREL. As I emphasized in the question, I do not need to teach the structural model. If someone knows the answer, please respond with a LISREL output.

Arash,

1) I did indeed work with LISEL for several years

2) I exactly answered your question: The loading is fixed to 1 (as the marker variable) - not estimated (hence, to test statistic).

I can understand that sometimes, threads drift a bit due to the dynamics. Your tone is a bit special, though. I am willing to help, remember that :)

All the best,

Holger

Dear Arash,

I can only agree with Holger, the first loadings of the endogenous LVs are not reported as they need to be fixed for identification purposes and thus making statistical inference is senseless. Similar is done for the exogenous LVs, however, for these the variance is fixed. One could theoretically also fix the variance of the endogenous LVs, however this requires non-linear model constraints and thus typically the first loading is fixed which is more convenient. Btw. this is not a specific LISREL 'problem'. this occurs in every software package for SEM as otherwise, the model is not identified.

Best regards,

Florian