Hello All,
I am currently using STATA 14 for mixlogit and wondering how to identify the case-specific or alternative specific variables to place in rand(). At the moment it is done arbitrarily. Can I get some advice on what to guide this process? I have read the paper by Arne Hole(2007) and in one example the case-specific variable(interacted with the alt) is not randomised. Is it only alternative specific variables that can be put in rand? Thank you
I am currently using STATA 14 for mixlogit and wondering how to identify the case-specific or alternative specific variables to place in rand(). At the moment it is done arbitrarily. Can I get some advice on what to guide this process? I have read the paper by Arne Hole(2007) and in one example the case-specific variable(interacted with the alt) is not randomised. Is it only alternative specific variables that can be put in rand? Thank you
