Analyzing MaxDiff Using Standard Logit Models

The Multinomial Logit Model and its various mixture generalizations are designed to predict a choice from amongst a set of alternatives. These models can be 'tricked' to analyze max-diff data by treating each set (i.e., block) in the MaxDiff experiment as two separate sets. In the first set the dependent variable is the Best choice. In the second set, the Worst choice is the dependent variable and the independent variables are multiplied by -1. Worked examples are presented in:

• Analyzing MaxDiff Using Standard Logit Models Using R
• MaxDiff Analysis Case Study Using R

See also

• Counts Analysis of MaxDiff Data.
• Analyzing MaxDiff Using the Rank-Ordered Logit Models With Ties.
• MaxDiff Analysis Case Study Using R.