From A-to-Z: Executing experimental methods with a focus on discrete choice experiments

Participation Prerequisites

- Basic knowledge in statistics

- It is not necessary that you have already collected data for your dissertation thesis. The course is designed for doctoral students at any stage of their PhD

Course Content

Day 1:

- Introduction to experimental methods and discrete choice experiments (Session 1 & 2) 

- Theoretical roots & experimental set-up decisions (Session 3)

- Hands on: Questionnaire design and survey setup (Session 4) 

Day 2: 

- Design properties and design generation techniques (Session 5)

- Estimation I: Basic techniques using Maximum Likelihood (Session 6) 

- Estimation II: Advanced techniques using hierarchical Bayes (Session 7)

- Hands on: Design generation & estimation (Session 8)

- Excurse: My two cents on publication strategy

Day 3:

- Test for validity, interpretation and counterfactual simulations (Session 9)

- Advanced topics (Session 10)

- Hands on: Completion of study-preparations (Session 11)

- Presentation: Defense of planned study (Session 12)

Relevance of the course:

Proper methods for measuring preferences are crucial for understanding and predicting decisions made by individuals, managers, organizations, and consumers. Only with this knowledge can one identify the most effective management actions. These methods include traditional discrete choice experiments, such as Choice-Based Conjoint, ranking-based methods like Best-Worst Scaling (Case 1) and MaxDiff, and Self-Explicated Methods. They prove indispensable in research aimed at tackling complex challenges across various disciplines. For instance, they are used to quantify consumers' trade-offs between product characteristics and price in Marketing, explore households' decisions regarding high-interest rates versus low-risk in Household Finance, and examine app users' compromises between capabilities and privacy concerns in Information Systems, to mention a few examples.

The challenge is that preference measurement methods require a deep understanding of decision-making theory, a strong statistical foundation, and access to the tools to support their implementation. Through this doctoral course, I will introduce doctoral students to the most crucial preference measurement methods.

More importantly, I will not only explain how these methods work but also challenge doctoral students to apply them directly to a research question of their choice, e.g., a question from their dissertation project or another area of interest.

This PhD course will equip researchers with the ability to design and implement their own preference measurement study. Over the course of three days, doctoral students will plan, conduct, and analyze a study in a prototype manner. They will also gain access to necessary tools and receive training on using them effectively.

Software:

During the course, you will apply the following software:

- DISE (http://www.dise-online.net/demo.aspx) for setting up the questionnaire

- XML Spy for setting up the questionnaire

- JMP for design generation

- Excel for estimation

- R for estimation

Prior experience with some of the software mentioned above is helpful, but not a pre-condition. I developed this course so that participants with no prior experience with any of the software can also follow.

Preparations:

Before the course, please prepare the following:

- Carefully select the topic of your own study. I highly recommend choosing a topic that fits with your own doctoral research.

- Work through the manual of how to set-up a questionnaire using the survey platform DISE (http://www.dise-online.net/demo.aspx). You can access the manual as PDF: http://www.dise-online.com/userupload/dise-manual-en.pdf

- Register for an account and create a test-survey on the sandbox-server:

-If you are not in the WHU network: http://www.dise.sandbox.whu.edu/dise-online

-If you are in the WHU network: http://192.168.160.5/dise-online (both are the same server)

The test-survey, should contain at least three pages, 1 radiobutton, 1 textbox, and 1 choice-set. Test your survey, whether everything works properly and access the results.

 About 2 weeks, before the course starts, I will provide a guidance on selecting a topic for your study.

Intended Learning Outcomes and Competencies

This course will be of interest to researchers from any discipline (e.g., Marketing, Entrepreneurship, Finance, and Information Systems) with no or limited experience in the application of experiments and DCEs in particular. It will enable participants to:

- Identify suitable experimental methods

- Gain a thorough understanding of the steps required for the planning and execution of experiments

- Gain hands-on experience on each step of setting up a discrete choice experiments

- Understand the theoretical basis

- Learn about design generation techniques, study set-up, and interpretation of results

- Ability to critically reflect methodological issues

Instruction Type

Präsenzstudium/E-Learning

Form of Examination

Beside attendance in this course, you must demonstrate your ability to conduct your own preference measurement study (either alone or in groups of two people) on the topic of your choice. I encourage you to carefully select the topic of your study by aligning it with your own doctoral research.

We will cover the most challenging parts – namely the careful planning of the study set-up and survey creation – together in class. The hands-on sessions at the end of the day ensure you know how to apply the contents discussed throughout the day. You also must present, defend and discuss your prototypical study at the end of the third day.

The submission type is the PowerPoint-presentation containing your study background, its experimental set-up, and its first results, ideally at the end of the third day.

Literature

- Jordan Louviere, David Hensher, Joffre Swait (2000), Stated choice methods: Analysis and application.

- Kenneth Train (2008), Discrete choice methods using simulations, second edition.

- Additional reading material will be provided in class