Jinhua Zhao posted on Wednesday, March 25, 2009 - 6:55 pm
Dear Professors, I am running a mode choice (nominal) model with both latent class and latent variables. I tried 1) mode choice model with Latent Class. It worked. 2) mode choice model with latent variable. It worked.
But when I combine both latent class and latent variable in the mode choice model. Mplus complains.
The testing Mplus code except is as follows: ===================== USEVARIABLES are QA1 QA3 QA6 QA11 QA16 QA21 MALE CarOne CarTwo ModeChng;
CLASSES = c(2);
NOMINAL ARE ModeChng; !mode choice
ANALYSIS: TYPE = MIXTURE;
%OVERALL% f_Conven BY QA1 QA3 QA6 QA11 QA16 QA21; ModeChng ON CarOne CarTwo f_Conven; c ON MALE;
%C#1% ModeChng ON f_Conven;
========================= Note: 1. f_Conven is latent variable "convenience" measured by 6 indicators; 2. ModeChng is the mode choice 3. I want to test the different slopes of f_conven on ModeChng across classes.
When I run this code, Mplus complains "*** ERROR The following MODEL statements are ignored: * Statements in the OVERALL class: MODECHNG#1 ON F_CONVEN MODECHNG#2 ON F_CONVEN * Statements in Class 1: MODECHNG#1 ON F_CONVEN MODECHNG#2 ON F_CONVEN