Message/Author 


Hi, I want to run a APIM with indistinguishable dyads. For this, I have to constrain the error terms to be equal. I know how to constrain means/intercepts and paths to be equal, but cannot find how to do this with error terms. Could you please give advice. Thank you very much, Manja 


If the two variables are y1 and y2, you say: y1 (1); y2 (1); 


Dear Linda, Thank you for your prompt response. I have one more question: As I understood correctly the term [x1 x2] (1); constrains the means of x1 and x2 to be equal. Can I also constrain the variances of the two variables to be equal? Thanks again, Manja 


What I showed above or y1 y2 (1); holds variances or residual variances equal. Variances are estimated for exogenous variables. Residual variances are estimated for endogenous variables. The specification is the same. 

Lisa Aschan posted on Wednesday, July 18, 2012  7:04 am



Hi, I am trying to constrain two residual errors to be equal but when I do what is suggested above I don't get fit indices and the the following message appears: THE MODEL ESTIMATION TERMINATED NORMALLY THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES COULD NOT BE COMPUTED. THE MODEL MAY NOT BE IDENTIFIED. CHECK YOUR MODEL. PROBLEM INVOLVING PARAMETER 42. Parameter 42 are the parameters in the psi table which I have tried to constrain. My syntax looks like this: USEVARIABLES are pw1 x7 x9 x1 x2 x3 x4 phouse x5 x10 x11 x6 x8; CATEGORICAL are x9 x1 x2 x3 x4 x5 x10 x11 ; MISSING are all (99) ; CLUSTER is phouse ; WEIGHT is pw1 ; Analysis: Type = complex ; parameterization = theta ; Model: F1 by x1* x2 x3 x4 x5; F1@1 ; x11 on F1; x11 on x10; x10 on F1; x10 x11 (1); F1 with x7; F1 with x8 ; F1 with x6 ; x7 with x6 @0; x8 with x7 ; x8 with x6 ; x5 with x4; x9 on F1 ; x9 on x6 ; x9 on x7 ; x9 on x8 ; x10 on x6 ; x10 on x8 ; x10 on x9 ; x11 on x9 ; x11 on x7 ; x11 on x8 ; I would very much appreciate any guidance you might be able to give. 


Please send the output and your license number to support@statmodel.com. 

Back to top 