Software Quality Market : Design and Field Study of a Prediction Market

Given the increasing competition in the software industry and the critical consequences of software errors, it has become important to develop methods for early forecasting of software quality. This research utilized a novel forecasting approach based on recent work in using prediction markets, which has proved useful in a variety of situations. In a prediction market for software quality, individual estimates from diverse project stakeholders such as project managers, developers, testers, and users were collected at various points in time during the project. Stakeholders reveal their estimates by trading in software quality contracts in a market designed for the task. Data were collected using a prediction market and a survey from a live software project at three different stages of completion (requirements, initial release and final release phases). Analyses of the data suggests that i) forecasts from the prediction market differ from forecasts from surveys and ii) prediction market forecasts are more accurate predictors of eventual quality than surveys. Thus, prediction markets can serve as an additional mechanism for aggregating stakeholder beliefs about software quality