Lee L Zwanziger

lzwanzig@erols.com

Home:

Lee L. Zwanziger

7438 Gadsby Square

Alexandria, VA 22315-5291

phone: 703/971-6234

lzwanzig@erols.com

Office:

Lee L. Zwanziger, Ph.D.

Senior Program Officer

Institute of Medicine

National Academies

2101 Constitution Avenue, N.W., FO-3084

Washington, DC 20418

phone: 202/334-3042

fax: 202/334-2685

email: lzwanzig@nas.edu

When our estimation of the value of science for developing solutions to social problems overtakes its value as a generator of understanding, the result is to diminish the ability and the availability of science for either purpose. Different types of values function in science at different levels. At the most core epistemic level, values such as simplicity influence theory development. At the most peripheral political level, values such as paying political debts can influence funding and infrastructure development. In between, however, there is a wide range of more subtle interplay of science and socio-political values, reflecting temporally, often politically, contingent assumptions that are then expressed---frozen---into law and policy necessarily governing our wider social life.

In considering values in science, or whether value-free science is possible, it is important to consider first what types of values are in question. When I mentioned this conference, for instance, people said, "Yes, bioethics is very important right now." It is indeed, but I will be concentrating on intellectual, not moral, values. This is only one of many distinctions of types of values. The introductory material of the conference makes this point nicely: epistemic values like truth and simplicity at the ideal level might be illustrated with examples including values such as the maximization of explanatory relations with neighboring fields, or values involved in the assessment of which aspects of a theoretical framework are most in need of, and most amenable to, experimental grounding. Non-epistemic values have also been intensively studied, including the role of class, sex, and power relations, and of cultural assumptions, in shaping theory (theory-laden observation is one angle, but so also is observation burdened by prejudice, see Silverstein and Auerbach, 1999).

The results, especially of theory- or prejudice-laden analyses of science, are genuine and pressing questions about whether value-free science is possible. But in this talk, I want to pursue some similar problems from a different starting point. The former questions, regarding whether there can be value-free science, involve a high level of abstraction, since they address "the science" almost analogous to "the good," that is, the ideal and pure essence of what is science itself untainted and undiluted by other concerns. I hope to leave aside for a few minutes the question of whether such a value-free science can occur and concentrate instead on the science we actually have in the U.S., how it is in fact influenced by political values, and by implication, how it is valuable in political discourse.

I argue in this paper that, even as we correctly value science and scientists for their contributions in the development of solutions to social problems, it is crucial that such use of science and scientists remain conceptually separate from scientific inquiry itself. The application of science to solve problems, if not balanced by pursuit of science for its own sake, can inhibit the development of scientific understanding. Scientific activity undertaken for a social purpose is an activity of problem solving, not actually an activity of scientific inquiry. Problem solving is heavily and multiply constrained, and many of the constraints come from non-scientific sources. To the extent that science is practiced in this context, then the pursuit of science is also constrained so that, taken to its limit, the result is that there would be no scientific inquiry that is even substantially---if never completely---free of political constraint. At the same time, an exclusive focus on the socially conscious use of science to inform the development of solutions to social problems can also inhibit the development of best possible policies. The federal government regularly seeks scientific advice, but in practice the question posed to these scientific advisers will have been formulated under political constraints, as is the selection of the advisers themselves. This means that we can have very good scientists deliberating about a very good question, but where the entire inquiry is constrained by mutually reinforcing biases of non-scientific origin. In conclusion then, the ideal of neutral scientific judgement will be further eroded. Although an ideal, it is a crucial check, indeed the only effective check, against the real and present risk of compounded biases. Scholars in institutions of higher learning are the natural choice for leaders in guarding and protecting such an ideal.

The distinction between pure and applied research is familiar and, while misleading if taken too far, it is a useful starting point. When the primary point of the research is to answer a question in a theoretical framework, to further the development of the whole theory and to generate understanding, that is usually called "pure" research, while research is "applied" when the point is to answer a question in order to do something. These are not crystalline, immutable categories, and results in either one can result in gains for the other.

When research is undertaken for an external purpose, e.g., to cure a disease, solve a social problem, or in general to build a better mousetrap, it is essentially done to solve a problem. The word "problem" and the problem-solution unit of course have a history (Laudan, 1977). Maybe we can find a more neutral word; meanwhile, let me just explain what I mean, using this word.

Problems provide the impetus behind the search in such externally directed research activity. The problem includes a question to be sure, but it also must include by implication a set of decisions about what ought to be the state of affairs, the direction in which we should go to get from the current to the desirable state of affairs, the arena in which that change should take place, and the type of action that should be undertaken to bring about the change and by whom. Notably, the search for a solution to a problem is not motivated by an internal demand of the theory to be explicated or the phenomena to be understood. The point is to figure out how to meet a human need, using whatever scientific results may seem to suit that purpose.

For example, suppose we are concerned about reducing the use of fossil fuel, and seek a design for a car to use an alternative fuel such as solar batteries. The possible solution set is determined most of all by the problem designation: that is, the need for transportation (but not re-zoning or re-scheduling), the need for a car (but not light rail), and the need for batteries (but not methane). Of course the possible solution is also determined by available resources like time, money, and personnel, and by the features of the product that will be required if it is to be acceptable (weight, speed, distance between recharges). The designation of the problem (transportation, car, solar batteries, disallowing alternatives) already incorporates many decisions about what sort of approach to take. Let me call these the input constraints. As we proceed under those input constraints, we have to consider many other factors as well, such as what can be spent in the development of a solution in money, time and personnel. Since these resource limitations affect the production of the solution, let me at least for now call these the output constraints.

The example above emphasizes engineering, but public policy is similar in that it is also about how to accomplish a particular objective under a particular set of constraints. Engineering usually deals with the material world and public policy deals with either the material or the social world, in a political context. But either way the approach requires that the analysts (1) identify a problem; (2) identify the constraints (money, time, but also the background assumptions determining the permissible approach), and then (3) solve the problem within the constraints. Accomplishing this is no trivial matter, and depends on having both an extensive knowledge of the type of problem and the effects of the constraints, and an aptitude for creative thinking about how to work with that problem and within those constraints.

While both engineering and public policy concentrate on problem solving, a tremendous difference lies in the criteria of success. An engineering project’s standard of success is more likely to be agreed upon in the design phase, while in public policy the criteria may not be explicit. In public policies there may be strong and very different views of what success would look like and by what criteria success ought to be assessed. In a vaccination program, for instance, there may be intense but unresolved questions about whether the program should accomplish universal vaccination, or provide universal opportunity for vaccination, associated with differences over the relative importance of individual or parental rights compared to the good of society. An attempt to fully specify the goal could result in open disagreements that would make the entire program impossible to pursue. In other words, the input constraints in engineering may be more explicitly acknowledged than is possible, or rather expedient, to specify during public policy development.

The designation of the problem already incorporates many decisions about what approach to take, and these decision themselves reflect values: they reflect how we as a society have chosen and continue to chose to live and operate. These are clearly not scientific conclusions, are not based on scientific expertise, and in fact, are not in the scientific realm at all.

The selection of problems to study can be done by many criteria, but is done with an ultimate intent to accomplish something, not primarily to discover or articulate basic principles. It may well be necessary to find those basic principles if they are to be later available to develop solutions to problems. The government supports basic research, though the degree of support has always been contentious (Smith, 1990). But the government cannot support research to an unlimited degree, so the basic research is often selected because it is in an area thought likely to contribute to later problem solving. We have examples of this, as in the "War on Cancer," when the relevance of research in almost any area of biology became almost a joke. The serious side of this situation is that when the selection of problems comes from social need, then any theory development that is dependent upon association with that problem is being driven by forces, and possibly in directions, irrelevant to the generation of understanding in the field of interest. So scientific activity undertaken in a problem-solving context is constrained and directed and ultimately limited by non-scientific interests, and to the extent that this activity is preferred, scientific inquiry to generate basic understanding of the natural world may go undone.

The relation of science to our federal government is complex and has been for as long as the existence of the government (Smith, 1990). The U.S. federal government has long sought to assure itself of access to sound scientific advice. The federal government does this both through specialized bureaus (e.g., President’s Office of Science and Technology Policy, or for Congress, the Congressional Research Service and the now disbanded Office of Technology Assessment), and through consultation with external scientific advisers convened to give advice on particular topics. There are many non-governmental organizations that also provide advice on specific questions, such as the National Academies (Cochrane, 1978). Governments might seek scientific advice either for the development of policy for science, scientists and related questions of infrastructure and funding. On the other hand, governments may also seek scientific advice for the development of any public policy that touches upon subject matter that has been treated scientifically. The relatively greater reliance of the U.S. government, compared to European governments, on science advisers in the latter case of general public policy development is an interesting feature of our history (Smith, 1992).

The scientific advisory system allows the government flexible access to experts currently working in the field, and also avoids long-term employment when specific expertise is needed only intermittently. The ideal model is that the government will ask cutting-edge scientists about the state of the science (relevant to some policy question). The scientists will then summarize it and return the results to the government, which then can take that information into account along with political, legal, regulatory, budgetary and other information, to develop sound policy.

The government's requests for scientific advice are related to policy decisions, so the response of the science advisers has very direct practical consequences, and the issues may involve many citizens and a lot of money and other resources belonging to those citizens. The formation of policy is not a purely scientific exercise: it involves many other types of considerations. Because there are multiple views, especially of the non-scientific values included in the policy question, the public and the government have tried to normalize and open the operation of scientific advisers to the federal government through regulation (Jasanoff, 1990). The regulatory vehicle is the 1972 Federal Advisory Committee Act (FACA) and implementing regulations and policies in the various departments and agencies, and since 1997, in the National Academies as well. The FACA is an important step; like any body of regulation, the FACA was an expression of a concern about a current state of affairs and a hope about a better state of affairs. But the FACA itself actually is just a set of procedures, which, like any system of rules, can be manipulated both consciously and unconsciously with results that can undermine the best of intentions.

The question that is presented to scientific advisers comes from a human need filtered through a political process, and then interpreted by staff members who have to make sure they get a response that is possible to incorporate into the agency’s mission and scope of action, or that can result in fulfilling a contract. The staff members may have an interest, recognized or not, in viewing the question a certain way in correspondence with their own political, religious, etc views---that is, the staff may contribute to or reinforce the existing input constraints. At the same time, the staff members certainly will have an interest in interpreting the question in such a way as to be doable with the available time and resources. And "doable" is a very restrictive output constraint on the deliberations!

When governmental or other organizations contract for an outside review or study, they typically tout the independence of the advice they get. What if the group that is developing the advice is biased in its composition? Everyone involved knows that the advice will determine or at least influence policy, and many if not all people have values in areas independent of their professional expertise. The influence of the advisers’ backgrounds on their advice is readily recognizable and well known. On the one hand, the problem to be deliberated must have already been defined and formulated so as to have already greatly constrained thinking. The formulation of the problems also constrains the choice of persons to deliberate in that area. But even after all these constraints are considered, there can still be a pool of potential candidates and the choice among them is further constrained aspects such as the likely ability to come to some consensus---in short, by the staff’s interpretation of the problem and interest in getting a timely solution in fulfillment of the agency’s need, or of the contract or commission. It is crucial to realize that the types of biases suggested here in no way indicate any lack of competence on the part of the advisers, nor generally are they associated with any lack of integrity. They are present and compounded as a result of a process, and must be checked by standards that are maintained independently of that process (see also Longino, p. 80-2, 1990). I have observed a spectrum including both committees whose members embraced the possibility of influencing the political outcomes through the advisory committee process, and others whose members recognized political undercurrents and consciously objected to shaping them with scientific language, or to allowing themselves to be used as instruments in the political process.

Mutually Reinforcing Biases

Without such checks, the result will be that the science informing the social question may not be a complete consideration (both because of the input constraints on the question and the complement of individuals deliberating). The social problem thereby addressed may suffer doubly, from the above, but also by having the aspects of the problem that are not scientific being dealt with by scientists who are not necessarily the only or the best individuals to do that. When a group of scientific advisers is presented with a public policy problem whose input constraints are either underdetermined or at least are not explicit, that group may be forced to choose some constraining assumptions of its own, just in order to proceed. But the group’s non-scientific decision making about assumptions then skews the results to reflect the (non-scientific) beliefs or values of the advisers, who will in turn have been selected in accord with the values of the staff and administration asking for advice, and this occurs with the special authority of scientific expertise.

In conclusion, I do not think that the ideal of completely neutral science, particularly within single individuals, is possible. The ideal is also a standard that can be approached to varying degrees. This approach depends on (1) taking seriously both science in its possibilities and within its limits, and democracy, so as to bring balance to policy making that is informed by science, and (2) on the individual commitment of scientists as scholars to explicate and maintain the distinction between scientific and extra-scientific matters. As with many other areas of our human effort at civilization, I do not think this is something we can fix, finish, and file. Rather, it is an ongoing continual effort that is truly well worth continuing. This is not an area where we can afford to stop with knowing and arch observations of politically laden conclusions—we all know this occurs and we ought as scholars to work with it as condition to be mitigated.

The leadership for continued vigilance to minimize politicization and to balance bias in scientific interpretation, not to mention to maximize democracy in public policy making, should come from those who are schooled in habits of thought that are both broad and rigorous. That should put both philosophy and science in the vanguard. But it would also require a shift in attention away from analysis of the political aspects of the policy process as though we were speaking of inevitable non-cognitive forces of nature, and toward an engagement with and active support of scholarly standards on the assumption that the policy process is human and so human participants can choose to do it to higher standards.

Cochrane, R. C., The National Academy of Sciences: The First Hundred Years, 1863 – 1963, The National Academy of Sciences, Washington, D. C., 1978.

Jasanoff, S. The Fifth Branch: Science Advisers as Policymakers, Harvard University Press, Cambridge, Massachusetts, 1990.

Laudan, L., Progress and Its Problems, University of California Press, Berkeley, California, 1977.

Longino, H. E., Science as Social Knowledge: Values and Objectivity in Scientific Inquiry, Princeton University Press, Princeton, New Jersey, 1990.

Olafson, F.A., "Where Ethics Has Gone." Academic Questions. Vol. 13(3): 31-38, 2000.

Silverstein, L.B., and Auerbach, C. F. "Deconstructing the Essential Father." American Psychologist: 54: 397-407, June 1999

Smith, B. L. R., American Science Policy Since World War II, The Brookings Institution, Washington, D. C., 1990.

Smith, B. L. R., The Advisers: Scientists in the Policy Process, The Brookings Institution, Washington, D. C., 1992.