Introduction
If you want to obtain higher precision for the most preferred items in your MaxDiff study Bandit MaxDiff is an adaptive approach that makes this easy to do. Rather than use a fixed list of items, Bandit MaxDiff uses a constructed (dynamic) list that is derived from your parent list of MaxDiff items. It is an adaptive approach that oversamples best items: Bandit MaxDiff learns from past respondents’ choices so it can oversample for the next respondents what are tending to be the most preferred items.
•Programming and Fielding a Bandit MaxDiff Study
•Sample Size for Bandit MaxDiff Studies
•Boosted Bandit: All Items Shown to Every Respondent
•Boosted Bandit: Respondents see a Subset of Items