Gemma Lewis posted on Thursday, December 18, 2014 - 3:33 am
Hello. I am trying to run an exploratory LCA on a dataset n=1070. I have 18 binary indicators. I started with 2 classes and continued until BIC deteriorated (7 classes). 6 classes has lowest BIC. However, for each model I get the error message below. I have tried using TECH10 to get bivariate fit but I am not sure if this generates a p value? 1. Are there any other solutions? 2. does this mean this is a poorly fitting model?
THE CHI-SQUARE TEST CANNOT BE COMPUTED BECAUSE THE FREQUENCY TABLE FOR THE LATENT CLASS INDICATOR MODEL PART IS TOO LARGE.
The chi-square test of the frequency table of categorical variables cannot be computed when the frequency table is too large. Usually it works best with 8 or fewer indicators. This does not say anything about the fit of your model.
TECH10 gives univariate, bivariate, and response pattern fit information. The standardized residuals are z-scores. Values greater than 1.96 suggest misfit.