Hello,

I have a quick question regarding the nested logit model. I am trying to develop a model as presented in the figure. It consists of two stages: First (Yes/No) and Second (A1/A2). However, I am concerned about the sample size in the second stage. In this case, does the small sample size in stage 2 affect the nested model's reliability? Is it better to develop two separate binary logit models? I would appreciate your help.
Thanks,
Woo

Array