Greetings, When estimating random parameter or latent class choice models, how can one indicate that there are varying numbers of choice exercises per individual? Forgive me if I've overlooked how to accomplish this in the NLOGIT documentation, but I cannot find it anywhere. Thanks, Richard