Meaningful & Reliable Information: Reasoning about Uncertainty and the Value of Information 🇬🇧 🇧🇷

In this session, you will learn how to place a value on data or information before obtaining it. For example, how much should you be willing to pay for a well test, a clinical trial, a COVID-19 test?

Value of information (VOI) is a fundamental concept within decision analysis and one of its most interesting applications. All decision professionals should have a firm grasp of the VOI concept, know how to calculate it, and understand its properties. This includes understanding many fallacies, such as the belief that the value of information is the greatest when we are the most uncertain.

We will begin by understanding how we can and should learn from data. That is, how can we move (inversely) from an observation (e.g., a medical test result) to estimate the probability of the underlying cause (e.g., the presence of a pathogen)?

Once we understand how to reason from observations, the next step is to use that information to make better decisions. To accomplish this, we will introduce the necessary decision analysis concepts of decision trees and risk aversion. With these tools, we will be able to place a value being able to act with perfect foresight (i.e., the value of perfect information), which places an upper limit on all information gathering activities. We will then discuss how to modify our analysis when the information it not perfect. With this knowledge, we will dispel many widely held, but incorrect, beliefs regarding the properties of VOI.

We will illustrate the tools and concepts discussed during the workshop using examples taken from a variety of industries.