How Not to Make Science Policy

Friday, July 30, 2004

“How can you tell whether a whale is a mammal or a fish?” a teacher asks her third-grade class. “Take a vote?” pipes up one of the pupils.

This idea might be amusing coming from a child, but it’s a lot less funny when applied by governments to the formulation of complex policies that involve science and technology. And it’s an approach that is becoming increasingly common around the world.

During the past two decades, the convening of citizens’ consensus conferences on a variety of issues first gained popularity in Denmark, where it is believed that non-experts “bring to the conferences a basic ‘common sense’ derived from worries, visions, general view and actual everyday experience as their basis for asking a number of essential questions concerned with the given subject.” Denmark has applied this approach to a broad spectrum of scientific and technological issues, including food irradiation; the new biotechnology (also known as gene splicing, or GM, for genetically modified) applied to agriculture, animals, food, and manufacturing; setting limits on chemicals in the environment; fishing policy; and human genome mapping.

It has metastasized widely. Britons had their say during the summer of 2003. To gauge public opinion in advance of a decision on whether to allow commercial planting of gene-spliced crops, at great expense the British government sponsored a series of public discussions around the country, as well as using more conventional methods, such as focus groups. Local authorities held scores of additional public meetings on the subject.

The head of the British debates’ organizing committee, Professor Malcolm Grant, called them a “unique experiment to find out what ordinary people really think once they’ve heard all the arguments.” But the reality argues otherwise. Mark Henderson, science correspondent for The Times (London), offered this view of the U.K.’s half-million pound white elephant: “The exercise has been farce from start to finish. I’m not sure I want the man in the street to set Britain’s science, technology and agriculture policy. One of the six meetings . . . spent much of its time discussing whether the SARS virus might come from [gene-spliced] cotton in China. It’s more likely to have come from outer space.” Henderson went on to say that the meetings were dominated by anti-technology zealots, the only faction that was well enough organized and cared enough about the issue to attend. This comports with reports that as many as 79 percent of the 37,000 questionnaire responses were orchestrated by activists.

Jan Bowman attended three of the events—including one in Stourbridge, “where both invited speakers opposed biotech”—and offered an assessment similar to Henderson’s. “At all of them the audience numbered no more than 60, and was overwhelmingly middle class, white, and already anti-biotech.”

The urge not only to sample but to respond to public opinion flourishes across the Atlantic as well. The National Science Foundation, whose primary mission is to support laboratory research across many disciplines, is funding a series of “citizens’ technology forums,” at which average, previously uninformed Americans come together to solve a thorny question of technology policy. According to the NSF’s abstract of the project, being carried out by researchers at North Carolina State University under a 2002 grant, participants “receive information about that issue from a range of content-area experts, experts on social implications of science and technology, and representatives of special interest groups.” This is supposed to enable them to reach consensus “and ultimately generate recommendations.”

The project, first funded in 2002 to support two panels and expanded subsequently under a continuing grant, calls for eight more panels of 15 citizens (who are “representative of the local population”) each. Their deliberations will be overseen by a research team, “composed of faculty in rhetoric of science, group decision-making, and political science,” that will test both “an innovative measure of democratic deliberation” and “also political science theory, by investigating relationships between gender, ethnicity, lower socioeconomic status and increases in efficacy and trust in regulators.”

At a time when federal budgets are under pressure and laboratory research funding is tight, the NSF has seen fit to spend almost half a million taxpayer dollars on this politically correct but dubious project.

Getting policy recommendations on an obscure and complex technical question from groups of citizen non-experts (who are recruited by newspaper ads) is sort of like going from your cardiologist’s office to a café, explaining to the waitress the therapeutic options for your chest pain, and asking her whether you should have the angioplasty or just take medication. (It might help, of course, if there were specialists in the rhetoric of science and in group decision-making having lunch at a nearby table.)

The first of these NSF-funded groups tackled regulatory policy toward agricultural biotechnology and recommended that the government tighten regulations for growing gene-spliced crops, including a new requirement that the foods from these crops be labeled to identify them for consumers. Both of these proposals are unwarranted, inappropriate, and contrary to the recommendations of experts, including those within the government and the scientific community. The output of the citizens’ panel illustrates that such undertakings have limitations in both theory and practice; non-experts are too often subject to their own prejudices and to the specific choice of materials and advocates to whom they are exposed.

Although involvement of the public is critical to its understanding of government policy, it is less useful for the formulation of policy, particularly when complex issues of science and technology are involved. Science is not democratic. The citizenry do not get to vote on whether a whale is a mammal or a fish, or on the temperature at which water boils, and legislatures cannot repeal the laws of nature. (However, on questions to which there is no scientifically “right” answer—such as at what age persons can vote or at which they should no longer drive or whether we should carry out more manned exploration of the moon—public opinion can play a critical role.)

Thus, one should be wary of the attempts in various countries to sample public opinion as a prelude to setting policy on biotechnology: In recent years the Netherlands, Japan, Norway, South Korea, Switzerland, France, Argentina, Denmark, and New Zealand, as well as the United States and the United Kingdom, have conducted similar exercises on biotech-related issues. Even if such opinion-sampling exercises were better organized, widely attended, and more representative, their purpose should not be to translate the vox populi into policy on subjects that depend on an understanding of the subtleties of science and technology. The nuances of statistics—the phenomenon of statistical “clusters,” for example—and of concepts such as “the dose makes the poison,” are not intuitively obvious. Permitting persons who do not really grasp the subtler aspects of quantitative issues to make policy is reckless. Even the efficient sampling of public opinion under such circumstances would founder on the principle that something worth doing at all is worth doing well.

The goal of policy formulation should be to get the right answers—in this case, that gene-splicing technology is essentially an extension, or refinement, of less-precise genetic techniques that have been around for at least half a century; that gene-spliced plants can make critical contributions to farmers, consumers, and the health of the natural environment; and that, except as science dictates in specific cases, the products of recombinant organisms should be regulated no differently than other, similar agricultural and food products.

There exists a particularly distasteful variation on the theme of a government permitting biased or ignorant public opinion to shape public opinion: Government officials disingenuously rigging the system in a way to make it appear as though the public demands a preordained and dubious outcome—one that benefits not the public but the bureaucrats themselves. This is what occurred during 2000–2001 when, responding to the demands of activists, FDA commissioner Jane Henney sought to create new, more stringent regulations for gene-spliced foods.

In 1992, the FDA had promulgated a scientifically defensible, risk-based policy, which said that the “regulatory status of a food, irrespective of the method by which it is developed, is dependent upon objective characteristics of the food and the intended use.” The policy reminded plant breeders and food producers that they had “an obligation under the [Federal Food, Drug and Cosmetics Act] to ensure that the foods they offer to consumers are safe and in compliance with applicable legal requirements.” However, it treated gene-spliced and other foods no differently and required case-by-case review by regulators only when the products raised specific safety concerns. Thus, the agency’s approach was consistent with the consensus of the scientific community regarding the regulation of gene-spliced products. This approach was widely applauded as regulation that made sense, relied on scientific principles, protected consumers, and permitted innovation.

The FDA policy defined certain potentially hazardous characteristics of new foods that, if present, required greater scrutiny by the agency, which could result in additional testing and labeling or exclusion from commerce. In other words, characteristics related to risk—not simply to the use of one technique or another—would trigger heightened regulatory scrutiny. According to the FDA’s 1992 announcement, such characteristics include the introduction of genes that code for proteins or mediate the synthesis of other added substances, such as fatty acids and carbohydrates, that differ substantially in structure or function from other substances typically found in the food supply. Heightened scrutiny by regulators would also be required if the genetic change altered a macronutrient (such as a new variety of citrus with diminished levels of vitamin C), caused a potent allergen to be present in a milieu in which a consumer would not expect it (a peanut allergen in a potato, for example), or enhanced levels of a natural toxin.

Thus, the FDA’s 1992 policy appeared to codify a risk-based approach to the oversight of new plant varieties. However, at the same time, and without the benefit of rule making or formal notification to industry, the agency created a “voluntary consultation procedure,” in which producers of gene-spliced plants were expected to consult with the agency before marketing their products. Without exception they did so. And thousands of food products that contain gene-spliced whole foods or ingredients have been regulated under the FDA’s formal 1992 policy and informal consultation procedure and now are found in American supermarkets. None has ever been shown to cause harm to human health.

The goal of activists, however, was not food safety but (stultifying) food regulation, and the FDA went along. The agency held a series of public meetings around the country—not unlike the U.K.’s “GM Nation?” with the important exception that American government officials had a hidden agenda. The FDA packed the panels with anti-biotech radicals, activists were permitted to stuff the ballot box at every stop, and—wonder of wonders!—the agency decided that public opinion favored stricter regulation of gene-spliced foods. Consequently, on January 17, 2001 (three days before the Bush administration took up the reins of government), the FDA published a proposal to make the voluntary consultation procedure mandatory and to establish prodigious data requirements—requirements that no conventional food could (or should) meet.

In the end, knowing that the new rule would not survive review by the Bush Office of Management and Budget, the FDA abandoned it, but the government’s chicanery in fabricating grassroots support for a wrong-headed policy is galling, nevertheless.

As the journal Nature editorialized a decade ago, “regulation of biotechnology products, whether in agriculture, medicine, pharmaceuticals or manufacturing, should be based on any inherent risk in the product, not on the process by which it is made.” As for critical decisions about medical interventions and the design of airplanes and bridges, the best insights are likely to come from experts.

The formulation of public policy toward science and technology can be difficult, to be sure, but if democracy must eventually take public opinion into account, good government must also discount heuristic errors and prejudices. The eighteenth-century Irish statesman and writer Edmund Burke emphasized the government’s responsibility to make such determinations. He observed that in republics, “Your Representative owes you, not only his industry, but his judgment; and he betrays, instead of serving you, if he sacrifices it to your opinion.”

In other words, although it may be useful, and also politic, for governments to sample opinion widely on high-profile public policy issues, after the consultations and deliberations have been completed government leaders are supposed to lead. Now there’s a novel idea.