It’s hard enough to make good decisions about relatively straightforward options like a career path or a move to a new location. But for industry and policymakers who make sweeping decisions that affect the environment--such as whether to drill for oil on protected lands or build new nuclear power plants--the financial stakes, potential consequences, and complexity of the choices defy a simple weighing of pros and cons. Such decisions often involve a baffling mix of cost, risk, benefits and uncertainty.

Into this complexity come decision scientists such as Douglas Crawford-Brown, a professor of environmental sciences, engineering, and public policy at the University of North Carolina (UNC), Chapel Hill, who uses his knowledge of science and mathematical modeling to help make sense of these problems. "I sit at the intersection of scientific modeling [and] public policy--looking at how those models are applied in ranking alternative policy solutions--and then philosophy, where I look at questions of how good the models are and when they become good enough to drive public-policy decisions," he says.

From theoretical physics, to scientific modeling, to decision analysis

Crawford-Brown migrated to decision science from theoretical physics, starting in the late 1970s. When the post-Sputnik boom produced a glut of young physicists, he saw limited opportunity for an academic career in that field. While he was working on his Ph.D. at Georgia Tech, members of the faculty encouraged him to develop another specialty tangentially related to physics, and he worked on scientific modeling. He finished his degree and left physics for good.

Crawford-Brown worked for a while on models of the effects of radiation and chemical exposure on human health, then wandered into decision analysis by the back door when the Environmental Protection Agency (EPA) started to use some of his scientific models. "I became really interested in the question of why anyone would use those models, considering I knew what all the problems were with them, and I was developing them as an abstract scientific activity," he says. He wasn't sure his models were good enough for policy decisions to be based on them.

"That then led me to a whole set of research areas that told me how you construct the ways that decisions are made so that people will understand the quality of the scientific information that's coming in. One of my concerns in the decision field is that people not just assemble information and then just make decisions on the basis of that information, but [also] that they come to understand exactly how reliable that information is, under what circumstances they can use it, and under what circumstances they should not be using it."

How good is the model?

Bob Clemen, a decision sciences professor at Duke University's Fuqua School of Business, emphasizes that the math, although important, is not the challenge in this field. "The math is not difficult. Even my students concede this,” he says. “My textbook is called Making Hard Decisions, and many of my students are delighted to call it Making Decisions Hard. But what's hard is having the mental discipline to think clearly about the decision situation itself, being very clear about what's uncertain."

Scientific modeling and decision analysis have roots in Bayesian statistics, an inferential statistics in which preliminary probabilities are assigned for certain outcomes ("priors"), then adjusted quantitatively in response to evidence, observations, and assumptions. Where classical statistics requires a direct measurement of each individual parameter, Bayesian statistics allows researchers to base an estimate on a variety of available scientific information. After analyzing the problem, scientific modelers take whatever information they have available and work them into a model.

It can be difficult to integrate different types of data with a weighting that properly accounts for degrees of uncertainty. One of Crawford-Brown's initial concerns with the use of his models was that theorists and experimentalists in the decision field rarely talked to each other, making the proper incorporation of data into complex models all the more difficult. "The physics perspective is that modeling and experimental work are intimately linked. You have to have both of those groups working together to really have the science functioning properly," he says.

Since then, he has built collaborations with epidemiologists and other field researchers in an effort to understand whether his models replicate what experimentalists observe in the environment. "Models are simplifications," he says, "and sometimes those simplifications can limit not only your conclusions but also the kinds of questions you might ask about an environmental problem or disease."

According to Crawford-Brown and Ken Reckhow, a colleague at the Nicholas School of the Environment and Earth Sciences at Duke University, thinking about how a model might be used, and its overall utility, is a critical difference between a decision scientist and an environmental modeler. An environmental modeler presents how the science works, but a decision scientist uses decision theory to take into account both the scientific information in a model and its uncertainty, and to help policymakers weigh that information against other factors, Crawford-Brown says. Humans don't deal well with either uncertainty or ambiguity, notes James Hammitt, director of the Harvard Center for Risk Analysis. One of the frontiers in this field is working to make that uncertainty as clear as possible for a decision-maker.

Combining science, modeling, and public policy

Going from a model to a public policy decision involves understanding the needs of the stakeholders and the context, says Clemen. Because policymakers often are unfamiliar with the environmental science involved, a decision analyst serves as a conduit to translate the data into a useful framework. In one such study, Melissa Kenney, a graduate student in the Duke University School of the Environment, is developing decision methods that translate standard water quality measurements into practical information that policymakers in North Carolina and other states can use to make decisions about natural resources.

Decision scientists also factor into their models issues such as social justice, economic concerns, and congressional mandates. "What often drives scientists nuts is that scientists come up with the answer based on the science, and then they're stunned that society didn't adopt their answer," Crawford-Brown says. Science isn't some sort of trump card in policy decisions, he says. "Science tells you what is. It doesn't tell you what ought to be," and that, he says, is an equal concern of the policy maker.

Career outlook in decision sciences

Environmental decision analysis is still emerging as a separate field, and most people use these skills as part of their larger job responsibilities. In the last few years, John D. Graham, former administrator of the Office of Management and Budget, has promoted the use of these methods in government agencies like the EPA. Crawford-Brown, Hammitt and Clemen all have observed an increase in interest in environmental decision analysis--and an increase in job opportunities for people with the right skill set. Many of these opportunities, they say, are with consulting firms or with the EPA.

However, although decision science has become a field in itself, the main importance of the field in an environmental context is that it augments the tool kit of the well-trained environmental scientist, they note. Whatever field of science you are pursuing, understanding how to analyze uncertainty is fundamental, says Crawford-Brown. It also works the other way: A strong foundation in a discipline, and an intimate knowledge of the science involved, is a key to making good decisions.

So the first step, says Crawford-Brown, is a solid foundation in a scientific discipline. From there, an aspiring decision scientist should pursue training in decision theory, and he recommends getting practical experience in how organizations make decisions in the real world. Employers looking at environmental issues don't hire candidates for their decision analysis skills alone; rather, they hire good scientists who also possess decision-analysis skills, Crawford-Brown and Hammitt agree.

Anecdotal evidence suggests excellent employment prospects. The ten decision-science graduates each year from UNC Chapel Hill's program are "snapped up" almost immediately, Crawford-Brown says. Approximately, one third of those graduates eventually move into academia, whereas the other two-thirds pursue careers in government or industry. A Ph.D. isn't necessary for careers in the government and the private sector, but environmental decision analysts with master's degrees find their opportunities for promotion can be limited without that degree.

In academia, decision scientists are often scattered in schools of business, public policy, and in engineering or operations management, where departments offer individual courses in decision making. Graduate degree programs exist at universities including Harvard, Carnegie Mellon, Duke, and UNC Chapel Hill. Many people first approach decision analysis mid-career, when a job project demands that they learn and apply formal decision making techniques. In these situations, they might get executive training through short courses to get the skills that they need.

Although environmental decision analysis offers the rewards of bringing together science and public policy, "it is a field of study that has a lot of political, economic, and legal landmines that you can step on," Crawford-Brown says. People should know that decision scientists can get caught in political and legal crossfire, he says. Approaching the field with self-confidence is a key to success.

This material is based upon work supported by the National Science Foundation Grant No. SES-0549096. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Sarah Webb has a Ph.D. in bioorganic chemistry. She writes from Jersey City, New Jersey.Comments, suggestions?

Please send your feedback to our editor.

Sarah Webb writes about science, health, and technology from Chattanooga, Tennessee.