A respondent-friendly method of ranking long lists
James Heyman
University of St. Thomas
John Sailors
University of Scranton
Introduction
Suppose a marketing manager wants to know consumer preferences for the various ice cream flavours her company produces. Inferring preferences from sales data is problematic because not all flavours have the same distribution intensity, shelf space in the stores that carry them, or promotional support. Modelling preferences by accounting for flavour dissimilarities is complicated and time consuming, and so the manager turns to surveying consumers. A contemporary approach would be to use adaptive-choice-based conjoint or maximum difference scaling, but these require...