Using prediction markets for decision support

Organisations have always faced the challenge of making decisions in the context of large, dynamic, complex environments. The information technology revolution has increased by several orders of magnitude the information available to decision makers. However, information technology offers the promise of not just improving the information available to the decision making process, but also transforming the decision making process itself. This dissertation focuses on a relatively novel form of group decision making called a prediction market. Prediction markets use a market mechanism to aggregate the information held individually by a large, geographically and temporally separate group. Prediction markets are proposed as useful tools in a wide variety of organisational decision making situations. They are optimally deployed using information technology, which allows the defining characteristics of prediction markets such as scalability to be utilized to maximum effect. This dissertation investigates prediction markets as group decision making tools, and evaluates their effectiveness and utility as organisational decision making tools. To this end, the literature on prediction markets and group decision making has been reviewed in depth, with a particular focus on integrating these two domains. A synthesis of these knowledge domains allows several novel insights to emerge. The literature review also drives the empirical data collection phase of the research project. Methodologically, the research strategy was a longitudinal study of a prediction market operating in a realistic context. A variety of research methods, including observation of trading behaviours and tests of behavioural traits were used to gather data to answer the research questions which emerged from the literature review. The theoretical and empirical work detailed in this dissertation makes two contributions. First, the empirical work presented in this dissertation explores the effect of prediction market participation on individuals. It demonstrates how individual’s skills and behaviours are affected by prediction market participation. Second, the empirical work presented also examines the impact that individual’s personality traits have on prediction market participation and performance. This dissertation provides empirical data supporting the contention that prediction markets are valid decision support systems. It shows that they are a valuable tool in the decision making armoury of the modern organisation. By demonstrating that prediction markets have a positive effect on the behaviours of individuals, it provides a further rationale for their deployment in modern organisations. It demonstrates that personality traits do not overwhelm other considerations mediating the performance and participation of individuals in prediction markets, and thus emphasises both their general utility and the importance of the careful design of incentives in optimizing prediction market performance.