Research

My primary research interest is process tracing. More particular I m interested in information search in risky decision making situations. This topic is fascinating both from a methodological and content perspective. On the methodological side I developed a research tool to conduct online information search experiments,called WebDiP (Web DecisIon Processes). On the content side my research currently focuses on process data for various phenomena like endowment or time discounting. In a longer followed interest the framing effect caught a lot of my attention. In the following paragraphs I will explain more about these primary and secondary research interests.

Research Methods and Online Research

Through my curiosity for computers in general and computer based research methods in particular, conducting research online offered an interesting new opportunity. Together with a computer scientist I developed an open source software package called WebDiP. This package enables are searcher to conduct online information search experiments without having programming knowledge.The first paper (Schulte-Mecklenbeck & Huber, 2003) resulting from WebDiP demonstrated that data collected on the Internet are comparable in quality with data collected in the laboratory. In a follow up paper (Schulte-Mecklenbeck & Neun, 2005) we recently gave a more general introduction to the usage of WebDiP to bring the program to a greater public.Together with Ryan O. Murphy (Columbia University, Center for the Decision Sciences) I am currently working on a paper introducing new possibilities for the analysis of sequential data.

This data type is typical in information search and process tracing experiments. However, the methods used to analyze such data is still relatively crude. Through the application of methods which we lend from computer science and biology (i.e., Markov Models) we should be able to add new perspective on the problem of dealing with highly dependent, sequential data.

Process Tracing

Quite general there can be two groups of models identified within the decision making literature: structural- and process-models. Structural models have an input and an output. The main issue of such models is the analysis of the second part – the post-decisional choice. Using these models it is however difficult to gain insight into the processes going on during a task. Process models (methods) overcome this drawback by observing traces a decision maker leaves, actively or passively, while accomplishing a task. The order, the amount of acquired information, the timespent on certain information items as well as the causal representation of the decision task are of interest.

In Switzerland I worked together with Oswald Huber on a computerized method for his Active Information Search (AIS) paradigm. At the moment I am fortunate to be able to work with Eric J. Johnson one of the pioneers in process tracing research. Using different process tracing methods like MouselabWeb or WebDiP we investigate phenomena like endowment or discounting (both in consumer decision making tasks).

The framing effect is a well studied phenomenon in the decision making literature. However, most of the conducted studies do not meet an important criterion of experimental economics:subjects should always be payed for participating in an experiment.The reasoning for this is the increase i effort and seriousness in an experiment. A second problem evolving from the usage of mainly hypothetical rewards is the transfer from the laboratory to the real word setting. Whether phenomena hold when real money is at stake or real effort was invested is an important question. We (Kühberger, Schulte-Mecklenbeck & Perner, 2002) were able to demonstrate that neither the quality of the amount to win (hypothetical versus real) nor the size of the amount (small versus (relatively) large) had an influence on the framing phenomenon.An interesting feature of this experiment was that participants were asked to ‘work’ on a different task first and got payment for completion. Following this the framing manipulation was introduced.So not only real money was at stake but also their already invested time.

More recently I was running experiments together with Anton Kuehberger on the underlying processes when participants search for information in framing tasks with different setups (2 options, 4 options or 6 options).