TU Delft
print this page print this page     
2017/2018 Technology, Policy and Management Master Complex Systems Engineering and Management
Statistical Analysis of Choice Behaviour
Module Manager
Name E-mail
Prof.dr.ir. C.G. Chorus    C.G.Chorus@tudelft.nl
Name E-mail
Prof.dr.ir. C.G. Chorus    C.G.Chorus@tudelft.nl
Dr. E.J.E. Molin    E.J.E.Molin@tudelft.nl
Responsible for assignments
Name E-mail
Prof.dr.ir. C.G. Chorus    C.G.Chorus@tudelft.nl
Co-responsible for assignments
Name E-mail
Dr. E.J.E. Molin    E.J.E.Molin@tudelft.nl
Contact Hours / Week x/x/x/x
Education Period
Start Education
Exam Period
Course Language
Course Contents
If one wants to design effective systems and policies, one needs to understand the behavior of users of systems, and how people respond to policies. This course sets out to provide students with in-depth knowledge and hands-on experience with statistical methods that have proved very effective in gaining such an understanding of people’s (choice) behaviour.

Importantly, this course takes a quantitative, statistical perspective on choice behaviour. Methods taught can be used to quantitatively analyse and forecast aggregate level phenomena – such as market shares for, and economic benefits of, new public transport services – based on sound individual level theories, supported by appropriate datasets and cutting edge statistical methods.

The models, methods and techniques that will be taught in this course, are applicable - and have been used - throughout a wide variety of domains, including but not limited to Transportation, Health, Marketing, Environment and Energy, and Political sciences. As such, knowledge of these methods prepares students optimally for a wide range of (quantitative, empirical) topics for their graduation project.

The course will cover both canonical (and Nobel Prize winning) methods, as well as more recently proposed ones. Both lecturers are highly active in the international research community, and their work forms part of the academic state of the art. For example, Random regret models developed by Chorus have been incorporated in several leading Econometric software packages, and are now taught in Econometrics courses in various corners of the world.
Study Goals
Students will acquire an in-depth knowledge concerning the below mentioned methods, as well as the skills to use these methods themselves. Finally, they will acquire an understanding of the usefulness – and limitations – of these methods, and an understanding of how the methods can and should be used in combination.

Knowledge transfer will take the form of a series of lectures, whereas skills will be developed in tailor made computer exercises. Specifically, during the computer exercises, students will learn how to design a statistically optimal choice experiment, and how to analyze data obtained from such experiments. In addition, they will learn how to interpret the outcomes of such analyses and how to use them to evaluate and design policies. Just to give one example: a clever combination of choice experiments and choice model estimations will help students to determine the societal benefits (in euros) of a technology that eliminates CO2-emissions in container transport.

- Choice behaviour modelling methods (Chorus)
o Random Utility Maximization & the Multinomial Logit model
o Mixed Logit models
o Random Regret Minimization

- Stated Choice data collection methods (Molin)
o Data-collection paradigms
o Stated choice experiments
o Efficient design approaches.
Education Method
Each method will we covered by means of:
o Lecture series (2 per week)
o Exercises (1 per week, in which concepts taught during the lectures will be used in hands on exercises)

What knowledge level do we expect students to have before starting the course:

- introductory knowledge in statistics, including among others, knowledge of basic terms and statistical procedures, such as variance, correlation, normal distribution, standard errors, t-test, reliability intervals, statistical testing of hypothesis.

- basic knowledge of survey data collection concepts and methods (sample, population, sampling, questionnaires)

In case of deficits, the student needs to invest additional time to acquire this knowledge. Suggestions for literature are provided, as well as a previously recorded lecture.
Examination: written exam, consisting of a combination of open and closed questions, and computational work.