Mai Sherif posted on Thursday, March 14, 2013 - 7:38 am
I am running a SEM for longitudinal data. My indicators are categorical and in order to achieve time invariance I set loadings and thresholds of the same items to be equal at different time points.
My question is: by setting thresholds of the categorical variables to be equal does this imply that thresholds of the underlying continuous variables are also set to be equal? what is the distribution of the underlying continuous variables...is it standard normal or do they have means and variances that are estimated?
The thresholds of the underlying latent response variable are used to discretize the latent response variable into the observed categorical variable. The latent response variable is normal with a mean and variance that are a function of the growth model. It is not standard normal.