Martijn Willemsen Martijn Willemsen researches the cognitive aspects of Human-Technology Interaction, with a strong focus on judgment and decision making in online environments. His applied research focuses on how online decisions can be supported by recommender systems, and includes domains such as movies, health related decisions and energy-saving measures. From a more theoretical perspective, he has a special interest in process tracing technologies to capture and analyze in detail information processing of decision makers.

Recommender systems

In this information overload society, many tools are available for helping people to make decisions among a multitude of options. These recommender systems employ smart algorithms that can predict how a person would like certain items, based on what he rated earlier, or what he watched (movies: netflix), listened to (last.fm), or bought before (amazon.com). Within our group, we study the psychological processes of decision making in these complex choice enviroments, focusing on how preferences should be represented in an algorithm, how we can prevent choice overload, and how we can enhance the usability of these systems. Martijn participated in the FP7 European project MyMedia.
On this project collaborating is established with Bart Knijnenburg, a former master student of Martijn, now doing a Ph.D. at UC Irvine. Their research has been focused on user-centric evaluation of recommender systems. In 2012, we published a paper in UMUAI on this topic.
Research research has focused on underlying psychological processes of recommender systems, such as choice overload and diversfication and memory effects in rating behavior.

Cognitive processes: process tracing and modeling

Another research direction is the use of process tracing techniques to get more insight into the cognitive processes underlying decision making phenomena, which is joint work with Eric Johnson (Columbia) and Ulf Böckenholt (McGill University). The project tries to get deeper insight into the cognitive processes underlying the decision making process by analysing and modeling process data using modern techniques for data representation (Icon Graphs) and statistical analysis (multi-level modeling). Currently a paper is published in Psycholological Review that uses this new approach in a comment to the Priority Heuristic, and in March 2011, our second paper on processes underlying loss aversion was accepted and published online in Journal of Experimental Psychology: general. Martijn and Eric are currently finalizing a third paper on context effects and process tracing.

Veni Grant: positive and negative descriptions and online reviews

From 2003 till 2006, a research program was executed (funded by a Veni-grant from NWO (Dutch Science Foundation)) that extended this research by investigating the way in which people describe choice options, in terms of positive and negative charaterizations. Within this project we investigated what type of descriptions people prefer to provide to others, prefer to receive themselves and what type they actually use in the decision process. Part of this research project looked into how people use online (customer) reviews in their decision making process while shopping online.

Comprehension checks

Together with Gideon Keren, Martijn has been working on an paper that shows that comprehension checks are important when performing experiments. We shows that the uncertainty effect, as found by Gneezy, List and Wu (2006) might have been caused by miscomprehension. This paper is now published in JBDM.

Dissertation: Explaining Asymmetries in Preference Elicitation

Martijn Willemsen started his Ph.D. under supervision of Gideon Keren in february 1998. Related to his Master-thesis, a Ph.D. project was formulated on the topic of the measurement of subjective values: The problem of preference elicitation. Research in decision making has shown that measurements of preferences are sensitive to both the context and the method by which the preference is elicited.
Explaining Asymmetries in Preference Elicitation: The Role of Negative Attributes in Judgment and Choice
The main focus of the project was on the asymmetrical role of negative (undesired) and positive (desired) attributes in choice and evaluation. Four methods of preference elicitation were studied in detail: choice, matching (making two options equally attractive), separate evaluation (rating of a single option) and joint evaluation (comparative ratings of two options). The thesis demonstrates that negative attributes loom more than positive attributes and that this negativity effect varies along different methods of elicitation. Negative attributes loom more in choice than in matching (Willemsen and Keren, 2002), and in separate than in joint evaluation (Willemsen and Keren, 2004). Within the thesis, the matching procedure was studied in more detail (Willemsen and Keren, 2003).

[click to download PDF version]

Web Experiments

Next to his scientific interests, Martijn has a work-related interest in designing and using web experiments. Besides the development of MouselabWEB, he is also working on an open source tool for participant administration and experiment management.