Hi,

I'm using mlogit to model a McFadden type multinomial logit model for choice of mode of travel. I have two different questions.
1- How can I define a nested multinomial logit model with mlogit or any other module?
2- I'm interested to understand how the model deals with the choice sets. in order to make mlogit to work, I need to add a row for each possible option. However in this case I'm trying to model mode choice for different trip purposes for activity based decision schemes. This will for instant mean that if you leave your home toward work using your bicycle, your choice set for next trip, say work to gym can not include private car as it won't be available. I can in that case either let the variables for the private car dedicates option remain empty or somehow define a dummy, informing the model that the car option will not be available.
wHat I wonder is if I do so (leave the non-available option row empty, whether mlogit or any other module will consider the specific person or will it like in the case of regression modelling, exclude the person totally?