Introduction to prediction and the value of information

Abstract

Predicting the consequences of alternative actions in terms of the objectives is central to decision making. Modeling in the broadest sense, from simple to complex and based on data or expert judgment, comprises the essential toolkit for making decision-relevant predictions. Gaps in knowledge and the resulting uncertainty can make predictive modeling challenging. Gathering information to address knowledge gaps, thereby reducing uncertainty, can improve predictions. However, within a decision analysis, the value of information gathering depends on the extent that reduced uncertainty will improve the decision’s outcome. Decision makers commonly confront the choice to proceed directly to a decision in the face of uncertainty or to delay and attempt to reduce the uncertainty significantly before making the decision. Value of information analysis can help make a smart choice. This chapter introduces the purpose, approaches, and tools for addressing knowledge gaps within decision analysis. The three case studies, which follow, illustrate some of the challenges and solutions encountered when addressing knowledge gaps within a decision analysis.