Understanding and modelling heterogeneity of human preferences for engineering design

Christopher Hoyle, Wei Chen*, Nanxin Wang, Gianna Gomez-Levi

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

25 Scopus citations


In today's competitive market, it is essential for companies to provide products which not only achieve high performance, but also appeal to the tastes of consumers. Therefore, a key element of design is an understanding of consumer preferences for product features. In this work, the random-effects ordered logit model is proposed as the modelling framework to capture the impact of both product and human attributes on consumers' ratings of qualitative system and sub-system attributes. To support the modelling, a series of methodologies are developed to both understand and model the influence of consumer heterogeneity upon product preferences. To illustrate the methodologies, a human appraisal experiment for understanding preferences for automobile occupant package design is analysed. An issue with analysing human appraisal experiments is that the effect of respondent heterogeneity must be understood to separate the influence of design factors from that of human factors. Hierarchical Bayes estimation and cluster analysis are used to gain an understanding of respondent rating styles, which are subsequently modelled explicitly in the ordered logit model. Smoothing spline regression is used to determine the functional form of the ordered logit model. The proposed ordered logit model is validated using a vehicle design case study.

Original languageEnglish (US)
Pages (from-to)583-601
Number of pages19
JournalJournal of Engineering Design
Issue number8
StatePublished - Aug 2011


  • Consumer preference
  • Heterogeneity
  • Human appraisal
  • Ordered logit
  • Vehicle design

ASJC Scopus subject areas

  • General Engineering


Dive into the research topics of 'Understanding and modelling heterogeneity of human preferences for engineering design'. Together they form a unique fingerprint.

Cite this