Hi, I have a question concerning the model fit. I estimated latent linear growth curve models with three measurement time points. We estimated the effect of the variable comparison groups controlling for some variables, that is, for possible selection effects: individual and family background variables. The clustering of the data was taken into account with the calculation of the standard error by using the “TYPE=COMPLEX” option. The statistical method used for model estimation and model fitting was maximum likelihood with robust standard errors (MLR); for dealing with missing values I used full information maximum likelihood (FIML) estimation. I got a excellent model fit: Chi2= 4.16; df= 6; CFI= 1.000; RMSEA= .000. This is the case for the most of the models with this dependent variable. Is it often the case to get such model fits? Or is here something wrong? Thank you for your answer!