Here are seven paragraphs from the conclusion to Alan Mazur’s excellent book True Warnings and False Alarms: Evaluating Fears about the Health Risks of Technology, 1948-1971 (Resources for the Future, 2004, pp. 107-109, buy a copy here)– the concluding subsection is titled “Does the Truth Matter?” .

Mazur distinguishes between a “knowledge model” and a “politics model” for understanding public debates involving science. These distinctions are somewhat (but not entirely) related to the concepts of the “linear model” and “interest group pluralism” that I discuss in my forthcoming book, which is really about how to reconcile the fact that there are elements of both models in the reality of decision making. Neither of Mazur’s models accurately describes how the world works, we need both. Some of the more useful debates and discussions following my testimony his week reflected a paradigm clash between those who view the world through the lens pf the “knowledge model” and those – like me – who accept that the “politics model” also reflects some fundamental realities as well. Here is the excerpt:

In a democracy, the people or their representatives are free to spend public money as they see fit. Interest groups compete to channel funding to their favorite causes. If U.S. society chooses to allot far more money to cleaning up toxic waste sites, which harm few people, than to prevent teenagers from smoking, which creates an enormous health burden, that is our privilege as a nation.

Still, many risk analysts are disturbed when we fail to maximize the number of lives save per dollar of risk remediation. They point out that actions taken by government to avoid the consequences of an alleged hazard are often unrelated to the severity or scientific validity of the hazard (EPA 1982; Breyer 1993; Graham and Wiener 1995; Mazur 1998). The inference is that policy should be better aligned with science, and that irrational or inefficient elements of policymaking should be eliminated (but see Mazur 1995 and Driesen 2001 for limitations on this positions).

Yet public policy does not always flow directly from scientific knowledge. A value-laden subject decision is always involved, one that requires weighing pros and cons, costs and benefits, winners and losers. A wise policy choice for one party with certain interest may not be the wisest choice for a party with different interests. These considerations raise a question: does scientific evaluation of a warning matter at all?

Essentially two models show how science is applied to public policy. The first – call it the “knowledge model” – assumes that scientists can obtain approximately true answers to their research questions with methods that are fairly objective. This knowledge is used to inform public policy. For example, scientists can determine the health risks from exposure to fluoride at levels adequate to prevent cavities. Policy makers then use this finding as one factor in deciding whether to add fluoride to community drinking water. Such decisions cannot follow from facts alone, but facts ought to influence outcomes. If health risk is high, that should help shift the decision against fluoridation; if low, that should encourage fluoridation. The model makes no sense to anyone who denies that science can find correct answers.

The second model – the “politics model” – can be applied whatever one’s view concerning the objectivity of science. Here partisans use scientific findings as political capital to sway policy in the direction they prefer. If such partisans favor fluoridation, they will claim there is little health risk; if they oppose it, they claim a high health risk. I makes no difference if the findings are correct, objective, or honest as long as they are persuasive. The actors bury findings that work against their position, or attack them as invalid or inapplicable. In the politics model, scientific claims are used polemically, just like any other kind of political argumentation (Mazur 1998; Brown 1991).

The politics model has many proponents. Partisans in a particular controversy often see their goal as sufficiently important to justify any interpretation of scientific data that is favorable to their cause. During breaks from writing this final chapter, I am reading John McPhee’s (1971) laudable biography of David Brower, a major environmentalist of the postwar period. McPhee repeatedly describes Brower’s habit of making up “facts” to support his arguments against industrialists and developers. The biographer seems to regard this as an endearing tactic of the “archdruid” in his advocacy for wilderness preservation. Like McPhee, we sympathize with those who fight the good fight, accepting their argumentation when in other contexts it would be vexing.

But the politics model loses its appeal if applied to the entire array of technical controversies affecting policy. Science that is sufficiently malleable to serve any position in one controversy can serve any position in all controversies, and in that event science does not matter at all. The famous parable of “the tragedy of the commons” tells how each shepherd maximized his own herd’s grazing on the village green until no grass remained for anyone (Hardin 1968). In the same way, if each technical expert interprets data for his or her own convenience, with no attempt at objectivity, there will be no experts left with unimpeachable credibility, and we will all suffer for it. [emphasis added]