Why government isn't the best place to look for unbiased science
"The Bush administration is, to an unprecedented degree, distorting and manipulating the science meant to assist the formation and implementation of policy," declared a much ballyhooed report, Scientific Integrity in Policymaking, issued last week by the Union of Concerned Scientists (UCS)—a group not so much purely scientific as dedicatedly liberal-activist. The report was supported by an accompanying letter signed by 60 leading scientists, including 20 Nobel laureates.
The UCS report offers a litany of alleged abuses of science by the Bush administration, including ignoring the conclusions of a National Academy of Sciences National Research Council report on the scientific evidence for climate change; suppressing a chapter on climate change from the Environmental Protection Agency's annual report on air pollution; censoring EPA information on air quality; distorting information about reproductive health issues; suppressing information about the danger of airborne bacteria near hog confinement operations; manipulating science concerning the Endangered Species Act and forest management; and appointing unqualified and industry-linked scientists to federal scientific advisory panels.
For the sake of argument, let's assume that the Bush administration has done all that UCS accuses it of doing. This problem is not particular to Republican administrations—the very linkage of government and science almost guarantees some chicanery. Let's recall the halcyon days of the Clinton administration. In 1993, Princeton University physicist William Happer was fired from the Department of Energy because he disagreed with Vice President Al Gore's views on stratospheric ozone depletion. In 1994, President Bill Clinton rejected the finding from the Embryo Research Panel of the National Institutes of Health which declared that the intentional creation of human embryos for genetic research was ethical. Clinton simply banned any federal funding for such research.
And in 1993, the EPA used a meta-analysis of a number of studies to find that second-hand smoke caused lung cancer in adult non-smokers and serious respiratory problems in children. That may well be, but the EPA had to put its thumb on the scales in order to get the result it wanted. The agency included just 11 out of 30 known studies on second-hand smoke in its meta-analysis, and even then found no increased risk to non-smokers at the 95 percent confidence level that had been the traditional agency standard. So the agency simply moved the confidence level from 95 percent to 90 percent in order to get the result it wanted.
At the time, I talked to a member of the EPA's scientific advisory board, an epidemiologist working at a leading east coast university who requested anonymity. He told me that he knew it was inadvisable to change the confidence level. He didn't oppose the change, though, because he was afraid he would be kicked off the board if he didn't go along. "I wanted to remain relevant to the policy process," he explained. He was also an EPA grant recipient.
The UCS report did not cite a single instance of where science was "abused" in furtherance of a policy that it favored. To its mind only political conservatives misuse science. But it's not that quite that simple. Consider two current cases pending before the Food and Drug Administration (FDA).
The FDA has delayed approving Plan B contraception even after its scientific advisory board (SAB) voted 24 to 3 that it could be sold safely as an over-the-counter medication. Instead of adopting the SAB's recommendation as it normally does, the FDA decided to delay approval for 90 days. Why? Because some conservatives in Congress are afraid that it will encourage teenage promiscuity. (The same argument could be used to argue for a ban on condoms.) This shows how the FDA caves in to conservative political pressure, right? Well, the agency also recently overruled an SAB vote of 9 to 6 declaring that silicone breast implants were safe—this overruling at least in part because of pressure exerted by liberal women's groups.
Such politicization of science, in response to ideological pressue from all directions, is perhaps an inevitable result of government funding of science. And it is dangerous. As the old adage says: Everyone's entitled to his own opinion, but not his own facts. As Jonathan Rauch outlines in his superb book Kindly Inquisitors: The New Attacks on Free Thought, liberal societies rest on three pillars: democracy, whereby we decide who gets to wield legitimate coercive power; capitalism, whereby we decide who gets what; and liberal science, whereby we decide what is true. In a liberal secular society, science is the one standard of truth that most citizens can agree on. Thus everyone tries to show that "science" supports his or her point of view, pet project, or preferred policy. This makes the kind of distortions UCS points out—as well as the kind of distortions it doesn't point out—inevitable when government funds science.
The UCS's recommendations are naïve, given the above realities of political science. It recommends that the president issue executive orders directing an end to efforts to distort science. It declares that Congress should require that all appointees to scientific panels meet high professional standards and pass laws that protect against the domination of such panels by individuals tied to entities with a vested interest at stake. It also wants Congress to guarantee public access to government scientific studies and the findings of scientific advisory panels.
Various steps have already been taken to improve the quality of the scientific information available to policy makers. For example, in 1997 the EPA set new clean air standards based largely on research done by Harvard epidemiologists. However, the researchers refused to release their data to other scientists for review. Making regulatory policy on the basis of secret science is a bit problematic to say the least.
This situation provoked Congress to pass the Federal Data Quality Act (FDQA) in 1999. This act directed the Office of Management and Budget (OMB) to "issue guidelines…that provide policy and procedural guidance to federal agencies for ensuring and maximizing the quality, objectivity, utility, and integrity of information (including statistical information) disseminated by federal agencies."
Last November, under the FDQA, the OMB issued proposed regulations for peer review and information quality of agency science. The proposed rules would require public access to all data used to make regulatory decisions, and would require that scientific advisors disclose possible conflicts of interest, including the fact that they are recipients of grants from interested federal agencies.
Still, some fear that creating an official peer-review process may actually shield regulatory agencies from legitimate and needed scrutiny, especially from courts asked to review the adequacy of rule making processes. As regulatory experts Brian Mannix from the Mercatus Institute and Jeff Kueter from the Marshall Institute note in their comments on the proposed OMB peer-review guidelines, "Whether they involve human medicines, endangered species, air and water quality, industrial chemicals, transportation safety, or consumer products, the stakes involved in federal regulatory decisions are enormous. OMB's periodic reports on the benefits and costs of federal regulation show that hundreds of billions of dollars turn on these decisions, but it is also true that lives are at stake, on both sides of the equation. It is important to get these decisions right, within the limits of human knowledge and analytical and policy development processes."
The UCS report has to admit, "No administration has been above inserting politics into science from time to time." In the end, the best we can probably do is to continue to muddle through the great society-wide peer-review process that encompasses the most biased activist pamphlets as well as the first drafts of history offered in our daily newspapers all the way through the formal academic peer-review process in our leading scientific journals. The UCS report is best thought of as just another input to that vast ongoing process of peer review.