How to differentiate the fit of the measurement model from regression effects in PROC CALIS

svh — Mon, 27 Mar 2023 18:41:02 GMT

I am using the LINEQS statement in PROC CALIS to estimate a structural model in which part of the model looks like the diagram in the attachment. In short, our conceptual framework suggests that x1, x2, and x3 are manifest variables for the latent variable Y1. Y1, in turn, has a direct effect on a different latent variable (Y2) whose manifest variables are not shown. My question is that when I fit a structural model in PROC CALIS, I do not know how to differentiate the standardized effects of the measurement model from the regression effects. All of the effects are summarized in the table in the output titled "Standardized Effects in Linear Equations", but how do I know if my model is not including Y2 as a manifest variable for of Y1? I cannot seem to find this addressed in the documentation.

/* Example of PROC CALIS code */
lineqs
x1 = L1 Y1 + E1,
x2 = L2 Y1+ E2,
x3 = L3 Y1 + E3,
Y2 = B1 Y1 + D1,

Re: How to differentiate the fit of the measurement model from regression effects in PROC CALIS

jiltao — Tue, 28 Mar 2023 13:27:24 GMT

Y2 is not a variable in your data set, and therefore will be treated as a latent variable, not a manifest variable in PROC CALIS.

topic Re: How to differentiate the fit of the measurement model from regression effects in PROC CALIS in Statistical Procedures

How to differentiate the fit of the measurement model from regression effects in PROC CALIS

Re: How to differentiate the fit of the measurement model from regression effects in PROC CALIS