The mean and variance given in the summary statistics in Mx output are completely different from what is calculated in SPSS, SAS, or Excel. What is wrong?

For dependent variables: Check whether the number of observations is right in Mx. If not, Mx is probably including the missing value code as a real value. Always state missing value codes in the Mx script before reading in the datafile. If the number of observations is the same: check the order of the variables and that the labels match the actual data.

For definition variables: since Mx does not handle two different values for missing values, summary statistics for definition variables do not agree with SPSS etc. This is okay, since Mx does not use the definition variable when the dependent variable of the corresponding case is missing.