flexible to inform risk-management decisions (NRC 2009), the importance of uncertainty characterization and analysis will only increase. It should be noted that the increasing importance of uncertainty analysis does not necessarily imply increasing sophistication of computational methods or even increasing necessity of quantitative uncertainty analysis. As discussed in Science and Decisions: Advancing Risk Assessment (NRC 2009), uncertainty analysis is a component to be planned for with the rest of an assessment, and a simple bounding analysis or qualitative elucidation of different types of uncertainties may be adequate if it shows that a given risk-management decision is robust compared with competing options (NRC 2009).

Consistent and holistic approaches are necessary for characterizing and recognizing uncertainty (in particular the various types of uncertainty, including unquantifiable systems-level uncertainties, indeterminacy, and ignorance). Such approaches would allow EPA to articulate the importance of uncertainty in light of pending decisions and not become paralyzed by the need for increasingly complex computational analysis. In addition, applying uncertainty analysis coherently in all EPA’s arenas would ensure that a policy or decision is both tenable and robust (van der Sluijs et al. 2008) and would ensure that uncertainty analysis is a means to an end and is designed with the end use in mind. Similarly, uncertainty analyses that are billed as comprehensive but omit key sources of uncertainty have the potential to be misleading or to lead to inappropriate decisions about research priorities and interventions. Finally, EPA would benefit from communicating uncertainty more effectively. Uncertainty is often mistakenly viewed as a negative form of knowledge, an indicator of poor-quality science (Funtowicz and Ravetz 1992). There is therefore a perception that acknowledging uncertainty can weaken agency authority by creating an image of the agency as unknowledgeable, by threatening the objectivity of “science-based” standards, and by making it more difficult to defend itself in the face of political and court challenges. However, reluctance to acknowledge uncertainty can lead EPA to rely on tools and methods that cannot provide timely answers, can push the agency to use point estimates to defend what are policy decisions (see Brickman et al. 1985), and runs counter to the value of uncertainty analysis in informing research and decision priorities.


The committee has described the important emerging environmental issues and complex challenges in Chapter 2 and the many types of emerging scientific information, tools, techniques, and technologies in Chapter 3 and Appendixes C and D. It is clear that if EPA is to meet those challenges and to make the greatest possible use of the new scientific tools, its problems will need to be approached from a systems perspective. Although improved science is important for EPA’s future, it is not sufficient for fully improving EPA’s capabilities for dealing with health and environmental challenges. Better economic analysis, policy ap-

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement