Dear Bengt and Linda: I ran an Mplus analysis that consisted of a two-class latent variable with binary indicators. I regressed the two classes on three explanatory variables without specifying any thresholds, when I look at the output I get the same slope and variance estimates for the 3 explanatory variables for class 1 and class 2. Is this what I should expect? If so, why? If you need more details or the syntax or I wasn't clear please let me know.