The Real-time Delphi method is an established forecasting tool in areas where the questions addresses expertise from different fields and/or where conflicting goals and values have to be considered. Based on the idea of the traditional Delphi method it offers experts a structured communication method free of “inter-personal” and hierarchical effects, and social pressure. However there still exits some challenging factors, such as the high drop-outs of experts during the rounds and also the selection of the experts.
In the thesis we want to evaluate if prediction markets can be used as a tool to select experts that can be later consulted in a Real-time Delphi study. We also want to examine the information flows between the two established forecasting tools and see if individual weaknesses can be balanced out.
The study will be conducted on the FAZ.NET-Orakel platform http://orakel.faz.net or in the laboratory, depending on the exact research question.
The aim of the thesis is to examine several research questions in the context of combining Prediction Markets and Real-time Delphi. The exact topic can be defined individually also according to personal preferences but will usually include experiment design and evaluation.
Exemplary questions may be are:
- Are the best participants in prediction markets really the experts?
- How to motivate traders on prediction markets to reveal their insider information?
- Can social elements on Real-time Delphi platforms reduce drop-outs?
- Is there an information flow of Delphi surveys into prediction markets?
… and many more.
Besides of a general interest in the topic, we demand at least basic skills in statistics. Experience in the design and conduction of experiments or the design of surveys are of advantage. Programming skills are not required, but if interested it is also possible to include programming tasks (mainly web-development and Java-based technologies).
Language: German or English
If you are interested or have any additional questions (or suggestions), feel free to contact: