hat most risk models overlook, Harris and Harper say, is that Native Americans, because of their lifestyles and cultures, are impacted differently by environmental contaminants than the general population. If those consequences are neglected, the hazard will not be assessed accurately, cleanup levels will not be set appropriately, and technologies selected will not be protective of tribal health and environment. Offering a practical approach for comprehensively analyzing Native American exposure scenarios, Harris and Harper have submitted a paper for publication in Risk Analysis. Reviewers have recognized the originality of the work, noting that "virtually nothing has been published on this topic." But almost as original and unique as the work itself is the collaboration that has given rise to the work and the challenges the work presents to conventionally trained scientists. Harris is a member of the Confederated Tribes of the Umatilla Indian Reservation and the natural/cultural resource coordinator of the tribe's Special Science and Resources Program. After working on a number of risk assessments, he realized that "comprehensive" was defined differently by those outside the tribe. He knew how difficult it was to communicate why and how the social, cultural, and spiritual impacts should he considered in an environmental assessment. Harper, a program manager for PNNL's Health Risk Assessment Department, knew that if she tried mixing sociocultural values with science she might well invite the criticism that she was working with less than "real" data and was no longer doing science at all. But their collaboration met a timely need. The U.S. Environmental Protection Agency, the U.S. Department of Energy, and the tribes in the Washington-Oregon area had reached a common conclusion that the Hanford Site risk models were deficient in addressing the factors that affected the tribes. Heightening the importance of the issue was President Clinton's signature on Executive Order 12898, which required that the impact of federal actions on peoples with subsistence diets be evaluated. With the help of DOE's Indian programs coordinator, Harper was able to visit several tribes. "DOE was not sending me there to tell them something. My goal was to genuinely do risk assessment in such a way that it incorporated the tribe's lifestyle and values, as opposed to trying to persuade them of some viewpoint." After several visits, she eventually met Harris. "Stuart's background was in the general sciences and geology and he had a good basis for understanding risk assessment principles. That was important because he, not I, was the one who had access to information about tribal lifestyles. I had to trust that Stuart knew risk assessment backwards and forwards and that he would know the questions to ask, relative to input parameters used in risk assessment," she added. The opportunity to work as a colleague and partner with a scientist outside the tribe is rare, Harris explained. "American Indian tribes have had a long history of interactions with anthropologists who would come in viewing us as objects of study. They would extract confidential information, go off and evaluate it by themselves using their value systems, and publish without permission, often to our disadvantage." Because of this history, mistrust is inevitable, Harper said. "It takes time and effort to build credibility. And you can't change your story to suit your client." The scientific community must rebuild its credibility with the tribes, Harper continued, and until then, conventional scientific and statistical tools may not be appropriate. The information, however, may be just as rigorous, systematic, and defensible, even if it's not fully quantitative. That may make their work less publishable in conventional journals, but because she and Harris are looking at a broader set of risk exposure scenarios, she believes the work they are doing is actually more accurate. "Integrating the individual pieces of a fragmented assessment, with the goal of understanding the impact on overall quality of life, is what we're going after and trying to capture," Harris said. That in turn can guide many things, including technology decision making. Contamination of media, habitat functions, human toxicity, ecological toxicity, economic well-being, community cohesiveness, cultural practices, and impact on future generations are all facets of a holistic risk assessment. The error, he thinks, comes in trying to understand the sum of a contaminant's impact by simply pointing to just one dimension, when all these aspects incorporate the complexity of what it means to be "at risk." But in moving toward a risk assessment model that considers cultural and societal values, Harper found herself reexamining her definition of science and how it is practiced. "Most of us are conventionally trained to think that science is supposed to be pure and objective, with the observer and the observed not linked in any way. We forget that scientists make personal choices about what to study, what is important or not to measure, and how much the results mean; and we make these choices for some reason that often turns out to be based on personal values, which are often derived from societal values. I think that science methods are much more culturally biased than we like to think. You often have to step back and broaden your perspective." "I've also had to set aside preconceived notions about what is a 'fact' and who is an 'expert.' You find yourself dealing with the questions of epistemology--how do you know you know?" The issue of values-based science, which has received a lot of attention of late, is quite relevant to their work, added Harper. "Some would say that values-based science is an oxymoron. I believe it should be just as objective, verifiable, repeatable, robust, and systematic as any science can be. It just recognizes that science can be done in service to values and it is cognizant that science methods may be culturally biased." Harris offers an example: "The reason that it's hard for us to apply many risk analysis methods is that they evaluate risk based on suburban human exposures but do not incorporate many of the factors that are unique to our lifestyle." Another aspect of values-based science that the collaboration has led Harper to contemplate is how information is weighed in decision-making. "For example, it is known that Native American cemeteries are contaminated at Hanford and that access to ancestral lands and resources is restricted. But what weight is assigned to 'access to ancestral lands' to reflect its importance in making risk and technology decisions?" Those issues are frequent topics of discussion between Harris and stake-holder participants in the Community Leaders Network. "When you take into account our small population, how we belong to and are a part of the land, and the fact that we are already home and can't move, you realize the magnitude of the multi-generational threat that environmental contaminants pose to our whole culture. If this fact is valued by risk assessors and risk managers, it leads you to gather information on longer-term impacts of environmental contamination. That in turn may lead you to develop a technology with performance specifications designed to do more thorough, eco-friendly, permanent cleanups." Collaborating with a tribal scientist on a holistic risk assessment model offers the opportunity to find win-win solutions, says Harper, even though it is sadly seen by many as an us-them, win-lose situation. She acknowledges that finding equitable solutions to environmental problems that involve such a complex mixture of peoples and issues will not be easy, but a major lesson she has learned from her collaboration with Harris is that ultimately one must be willing to listen and change. "This may be a different perspective, but it is definitely not just a matter of preference or perception--this is real. Science has to make sense and feel right."