As Statistica has some nice features and graphics capabilities (and I believe the latest version of Statistica is about to be released, which includes a nice VB interface) I'm thinking of switching to this software. Although I'm aware it will not have the functionality of S+, SAS etc, I am concerned it may lack the programming flexibility of these packages. I'm particularly interested in constructing finite mixture models such as zero inflated Poisson and INAR Poisson with longitudinal count data. How easy is this in Statistica? Do I have to program? If so, what algorithms are available for maximizing likelihoods (eg EM) ? Has anyone used Statistica in anger, or any general comments/suggestions to make? Many Thanks Dave Yearling Advanced Communication Research PP 13 Main Lab Block BT Research Labs Martlesham Heath Ipswich IP5 3RE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%