Daniel posted on Wednesday, August 24, 2005 - 3:49 pm
I ran a GGMM that required algorithm=integration. I also used the missing data option (i.e., type=mixture missing). Of course, I cannot get tech13, however, I'd like to use the multivariate skewness and kurtosis options to assess how well I replicate the observed values with my model. What do you suggest I do?