Sequential Logit Model

From Survey Analysis
Jump to: navigation, search

The sequential logit model is a generalization of the Multinomial Logit Model. The model works by 'exploding' a ranking into multiple separate choices. For example, if a respondent has ranked five alternatives as follows: C > A > B > D > E, the the sequential logit model recodes the data as multiple nested choice sets as follows:

  • C is chosen from amongst A, B, C, D and E.
  • A is chosen from amongst A, B, D and E.
  • B is chosen from amongst B, D and E.
  • D is chosen from amongst D and E.

Generalizations of this model allow for changes in preferences and variances over the course of the choice sets.

A particularly useful aspect of this model is that its parameters are interpreted in the same way as with a Multinomial Logit Model (e.g., they can be used to create a Choice Simulator).

Also known as

  • Exploded logit.
  • Rank-ordered logit model.

See also

A more up-to-date version of this content is on

Personal tools