Emil Coman posted on Sunday, May 17, 2009 - 11:56 am
Hi, I have a model with 2 dependent imaginary latent variables, a crosslag with only time 1 real data http://www.ncbi.nlm.nih.gov/pubmed/3816351. I have 8 indicators for X1, 3 for Y1, and X2 and Y2 are both imaginary latents-no indicators http://www.springerlink.com/content/25354767133541v7 I tried this, after consulting the forum for defining latents with formative indicators, phantoms, and induced variables. I get error message: THE MODEL MAY NOT BE IDENTIFIED. PROBLEM INVOLVING PARAMETER 12. 1. Where is parameter 12 indicated in output? 2. What else I am messing up here? I know parameters are related , Gollob & Reichardt said: phiX = rhoX0X *(sdX/sdX0) and phiY= rhoY0Y* (σY/σY0) – betaXY*rhoX0Y0* (sdY0/sdX0) -these are not in model yet. emil coman
MODEL: Strs1 BY Strw11- Strw18; Seff1 BY Sefw11- Sefw13; LStrs2 BY; LStrs2@.155; LStrs2 ON Strs1@.093(phiX); LSEff2 BY ; LSEff2 @.512; LSEff2 ON Seff1@.269(phiY); LSEff2 ON Strs1 (bXY); LStrs2 ON SEff1 (bYX); OUTPUT: TECH4 MODINDICES (ALL) STANDARDIZED;
If you ask for TECH1 in the OUTPUT command you will see which parameter is 12.
There must be some observed variable connected to the two phantom variables you define using the BY option. If you are trying to create formative indicators, see the Topic 1 course handout on the website.
Emil Coman posted on Monday, May 18, 2009 - 10:35 am
Thanks, sorry, to bother again; there are no formative/or reflective indicators for 2 variables, only X1 and Y1 have indicators (8 and 3), X2 and Y2 have none (are 'real' latents). it's pretty much X1->X2 and Y1->Y2, and I want the crosslags X1->Y2 and Y1->X2 estimated. I read a previous posting by Prof. B Muthen : on March 30, 2007 "Looks like you are doing formative indicator analysis which is fine as you have specified it, although all you have to say is action_i BY; to define the action_i factor (nothing needs to follow BY)."