New study reveals better model for predicting discrete outcomes in research.
The nested logit model is a useful tool for studying outcomes. There are two main versions, and one (RUMNL) is usually better. A new implementation called nlogitrum is introduced to help with this. Many examples show the differences between the two versions.