The Use of Discrete Choice Experiments and MaxDiff (Best-Worst) Scaling

A discrete choice experiment (DCE) is a method for evaluating the influences on consumer decisions and behaviour. In a DCE consumers make choices between alternatives each of which is described by a range of attributes. Analysis determines the importance and impact of the attributes on the decisions. When one attribute is price then the analysis estimates the consumers’ willingness-to-pay and price elasticity for the other attributes.

Level 2 - runs over 5 days

Cam Rungie's research examines methods for analyzing decisions made by consumers. Through an ARC Discovery grant, and with colleagues Len Coote and Jordan Louviere, he further developed the analysis of discrete choice experiments. The past 30 years has seen big contributions in the study and understanding of discrete choice. Building on this work, his developments bring to the field the same forms of statistical analysis widely used elsewhere in the social sciences such as regression and factor analysis. Apart from this developmental research, he has extensive experience in applied research, marketing research and econometrics. He is an accomplished and passionate teacher of quantitative topics and supervisor of postgraduate research students. He has knowledge both the theory and application of discrete choice methods.

About this course: 

The course will be based around specific examples of DCEs and case studies in health, non-market valuation, transport and marketing. An example might be to measure the importance of the attributes of a health service in consumers’ decision to use the service. Other examples might be to measure the $ value  consumers place on grape-variety and brand in the selection of wine, or more abstractly on there being fresh water in the Murray River.


At the completion of the course candidates will be able to plan, conduct, discuss and critique DCEs. Technically this will include:

  • attribute selection,
  • experimental design, sampling,
  • on-line survey tools,
  • data formatting, analysis and interpretation.


Sessions will cover

  • conditional logit,
  • latent class,
  • random coefficient,
  • Bayesian,
  • the 'none' option,
  • interactions,
  • alternative specific attributes,
  • willingness-to-pay,
  • elasticity
  • and preference simulation.

Basic training in computer and statistical packages for experimental design and analysis will be provided. The course introduces choices processes and the fundamentals of random utility theory. In a well-designed DCE the analysis is straight forward. Extensive modelling is not required. The more important skills are planning and interpretation. Overviews of more advanced discrete choice methods will be included, but detailed considerations will be left to other courses.

Course syllabus: 

Day 1
An introduction to a DCE. Case study. Candidates will design and implement a simple DCE, complete it as respondents, and analyse the results.


Day 2
Steps in the DCE Process. Case study. Coding and Data Manipulation. Conditional Logit. Likelihood ratio tests. Interactions.


Day 3
Software for experimental design, on-line surveys and analysis. Random coefficient, latent class and Bayesian models.

Day 4
Choice processes, random utility theory, preference simulation, maximum likelihood, and choice models.


Day 5
Steps in the process. Willingness-to-pay and elasticity. Advances in choice modelling.

Course format: 

This course is run in a classroom with participants using their own laptops.

Recommended Background: 

You should have an understanding of survey research and/or elementary statistics, hypothesis testing and data analysis such as in SPSS, equivalent to the syllabuses of Fundamentals of Survey Research and/or Fundamentals of Statistics.

Recommended Texts: 

The following reference books are recommended:

  • Hensher D.A., J. M. Rose and W.H. Greene, 2005, Applied Choice Analysis: A Primer, Cambridge University Press, New York.
  • Louviere J.J., D.A. Hensher and J Swaite Jr, 2000, Stated Choice Methods: Analysis and Applications, Cambridge University Press, New York.
  • Orme, B.K., 2013, Getting Started with Conjoint Analysis: Strategies for Product Design and Pricing Research, Third Edition, Research Publishers, Madison. 


Prior reading of Orme (2013) will be beneficial.


Q: Do I need any prior knowledge before taking this course

A: You should have an understanding of survey research and/or elementary statistics including hypothesis testing and data analysis.


Q: How advanced should I be in my own research?

A: It is better if you have some understanding of the data you wish to collect in your research but this is not essential.


Q: Should I do any prior reading?
A: Prior reading is certainly not required but of course can be useful. See the books listed under recommended texts and readings and in particular Orme (2013).


Q: Do I need my own laptop computer?
A: Yes, there will be in-class exercises and readings requiring the use of a computer.


Q: Should it be operating Windows?

A: Yes, other platforms are not supported.


Q: Do I need any specific software?
A: No, but you will need the usual software including Excel, Word, a pdf reader and a web browser.


Q: Will I need to down load software during the course?

A: Yes, so be sure you have the appropriate privileges for your laptop.

Participant feedback: 

A lot of time was spent in discussions and answering questions. This was very helpful and relevant. (summer 2016)


Extremely useful to have an expert answer your specific questions and give you a good frame for understanding theory. (Summer 2016)


Lecture notes, readings and an annotate bibliography will be provided. There is not a text book.