Topic: Discrete Choice Experimentation for Learning User Preferences

Speaker: Dr. William Fisher

Date: January 22 at 10:00 a.m. ET 

Abstract

The design of engineered products and services is a complex decision-making process where designers must often perform tradeoffs and select design features that align with user preferences. One way to simultaneously explore these tradeoffs and learn user preferences is by conducting a discrete choice experiment (DCE). In a DCE, users are presented with a sequence of questions, or choice sets, where in each choice set they are asked to pick their preferred option among a set of alternatives. The data collected from the DCE can then be used to estimate the relative importance of different design features and aid the designer in selecting a design that best aligns with user preferences.

    In this talk, we will begin by introducing DCEs, including their statistical design and analysis. Along the way, we will illustrate the statistical design and analysis concepts through an example involving laptop design. Then, we will present a case study on learning user preferences for tooltip text in JMP, a statistical software system, and demonstrate how incorporating demographic information into both the experimental design the analysis can allow us to model preference heterogeneity across different demographic groups, thus giving us the potential to tailor the tooltip to the user. Lastly, we will conclude with a brief overview of various interesting extensions and alternative preference learning approaches.

Speaker Bio

William Fisher is a Research Statistician Developer on the Design of Experiments and Reliability team at JMP Statistical Discovery LLC.  His work at JMP Statistical Discovery LLC focuses on doing research in experimental design and developing user-friendly software solutions to experimental design problems. Before joining JMP Statistical Discovery LLC, William was a student at Clemson University where he received his PhD in Mathematical Sciences in 2025.

Register Now