ApolloChoiceModelling forum

Hi,

I've got a follow-up question about the individual alphas. I'm still new to Apollo and I am in the process of building a MNL model to analyse a DCE with 12 choice tasks per participant.

I would like to include sociodemographic characteristics into the model as we expect them to have a major impact on the results. While I do understand the theory of estimating separate models for individuals with different characteristics, I can't work out how to adapt the code to be able to use it for MNL models.
My understanding is that, using "apollo_mnl", I can use the "rows" arguments of "mnl_settings" to restrict the analysis to individuals who are e.g. female. But how can the separate models be estimated together?

I would greatly appreciate any help and suggestions!

Thanks,
Elena

Elena

there are detailed examples in the manual (see e.g. for MNL) and on the website. You just include these effects in the utility functions, for example as interactions. The rows command is something quite different as it means certain rows in the data are not actually used

Stephane