If there is one thing that defines and limits our efforts to better understand extreme and rare events it is uncertainty. Uncertainty arises from both an imperfect understanding of the rare events and processes we wish to study (e.g., terrorism, natural hazards), and the imperfect, out-of-date, and incomplete data we must work with in order to try and understand these events and processes. No data are perfect. However, uncertainty is more than a technical "failing" of our data (e.g., measurement error); it arises, in part, because there are simply some things that are unknowable (Couclelis 2003) or, as Fischer (1999) articulates, may not be knowable with precision (i.e., inherent vagueness).1 Nevertheless, outside of academic circles one rarely sees maps, GIS databases, or visualization systems that acknowledge these fundamental limitations. This omission is problematic because, as MacEachren (1992, p.10) notes: In the early stages of scientific analysis or policy formulation, providing a way for analysts to assess uncertainty in the data they are exploring is critical to the perspectives they form and the approaches they decide to pursue. In the last 15 years, researchers in GIScience have made great advances in defining, measuring, modeling, and visualizing uncertainty and data quality (notably, Buttenfield, Clarke, Goodchild, MacEachren, Fischer, Beard, Ehlschlaeger, see references). Indeed, uncertainty has become a central issue in GIScience research with numerous conference sessions and journal articles devoted to the topic. Despite this sustained attention, a basic question that remains largely unanswered is whether displaying uncertainty helps users. In other words, does displaying uncertainty on maps fundamentally change the way people think and problem-solve and ultimately lead to better decisions? In this paper I will (1) argue why we need answers to these questions, (2) briefly review and synthesize relevant research findings to date, (3) define what constitutes "better decisions," and (4) outline how we might proceed from here.
@TechReport{ harrower:2003:RUBD,
author = {Mark Harrower},
title = {Representing Uncertainty: Does it Help People Make
Better Decisions?},
institution = {University Consortium for Geographic Information
Science, Geospatial Visualization and Knowledge
Discovery Workshop White Paper},
year = {2003},
}