Following my previous post on paradigm shifts and on how “normal science” occurs, I’d like to continue that with a discussion of scientific consensus. To put this in context, I’m partly motivated by the recent controversy about
Roger Pielke Jr., a professor of environmental studies at the University of Colorado Boulder, who is also currently a science writer for Nate Silver’s FiveThirtyEight website. (The controversy has been covered on Slate, Salon, and Huffington Post.) Silver’s work has been lauded for its data-driven analysis, but Pielke has been accused of misrepresenting data, selectively choosing data, and presenting misleading conclusions about climate change, for example about its effect on disaster occurrences and on the western drought.

This is also troubling in light of a recent article I read by Aklin & Urpelainen (2014), titled “Perceptions of scientific dissent undermine public support for environmental policy.” Based on an analysis of a survey of 1000 broadly selected Americans of age 18-65, they argue that “even small skeptical minorities can have large effects on the American public’s beliefs and preferences regarding environmental regulation.” (Incidentally, a book by Pielke is among their references.) If this is right, then we are left with the question about how to achieve consensus and inform public policy related to important environmental problems. As the authors note, it is not difficult for groups opposed to environmental regulation to confuse the public about the state of the scientific debate. Since it is difficult to win the debate in the media, a more promising strategy would be to increase awareness about the inherent uncertainties in scientific research so that the public does not expect unrealistically high degrees of consensus. (And that’s obviously what I’m trying to do here.)

Already a decade ago, the historian of science Naomi Oreskes (formerly a professor at UC San Diego) in a Science article analyzed nearly 1000 article abstracts about climate change over the previous decade and found that none disagreed explicitly with the notion of anthropogenic global warming–in other words, a consensus appears to have been reached. Not surprisingly, Pielke criticized this article a few months later. In her rebuttal, Oreskes made the point that, “Proxy debates about scientific uncertainty are a distraction from the real issue, which is how best to respond to the range of likely outcomes of global warming and how to maximize our ability to learn about the world we live in so as to be able to respond efficaciously. Denying science advances neither of those goals.”

The short answer to the question, “How do scientists reach a consensus?” is “They don’t.” Once a scientific field has moved beyond a period of transition, the overwhelming majority of scientists adopt at least the central tenets of a paradigm. But even then, there likely will be a few holdouts. The holdouts rarely turn out to be right, but their presence is useful because a healthy and democratic debate about the facts and their interpretation clarifies which aspects of the dominant paradigm are in need of further investigation. The stakes are higher, however, when scientific debate involves contentious issues related to public policy. In those situations, once a scientific consensus appears to be reached and once scientists are sufficiently certain about a particular issue, we want to be able to respond effectively in the short or long term with local, national, or international policies or regulations or moratoria, depending on what is called for. In the meantime, the debates can continue and the policies can be updated and improved.

Of course, it is not always straightforward to determine when a scientific consensus has been reached or when the scientific community is sufficiently certain about an issue. A relevant article here is that of Shwed & Bearman (2010), which was titled “The Temporal Structure of Scientific Consensus Formation.” They refer to “black boxing,” in which scientific consensus allows scientists to state something like “smoking causes cancer” without having to defend it, because it has become accepted by the consensus based on a body of research. Based on an analysis of citation networks, they show that areas considered by expert studies to have little rivalry have “flat” levels of modularity, while more controversial ones show much more modularity. “If consensus was obtained with fragile evidence, it will likely dissolve with growing interest, which is what happened at the onset of gravitational waves research.” But consensus about climate change was reached in the 1990s. Climate change skeptics (a label which may or may not apply to Pielke) and deniers can cultivate doubt in the short run, but they’ll likely find themselves ignored in the long run.

Finally, I want to make a more general point. I often talk about how science is messy and nonlinear, and that scientists are human beings with their own interests and who sometimes make mistakes. As stated by Steven Shapin (also formerly a professor at UC San Diego) in The Scientific Revolution, any account “that seeks to portray science as the contingent, diverse, and at times deeply problematic product of interested, morally concerned, historically situated people is likely to be read as criticism of science…Something is being criticized here: it is not science but some pervasive stories we tend to be told about science” (italics in original). Sometimes scientific debates aren’t 100% about logic and data and it’s never really possible to be 0% biased. But the scientific method is the most reliable and respected system we’ve got. (A few random people might disagree with that, but I think they’re wrong.)