I have a very similar question as the previous one. The only difference is that in my case, I used a dual response for my none option. When I calculate the choice preference for the holdout task (which has the same number of products presented as in random tasks) shall I include none in my estimation? If yes, what should be the weight? Thank you.