Skepticism and the scientific consensus

It figures.

Some of the most interesting questions and posts showed up right before Christmas, just the time when I didn’t have time to discuss and (hopefully) expand upon them. Neither, I’m guessing, did anyone else, which is unfortunate because this post was about an issue worth further discussion in the skeptical blogosphere. I’m talking about a post in which fellow ScienceBlogger Martin Rundkvist made this rather provocative observation about skepticism:

A discussion in the comments section of the recent Skeptics’ Circle reminded me of something I learned only after years in the skeptical movement.

A real skeptic always sides with scientific consensus.

This may sound really unsatisfying and self-contradictory at first. Isn’t skepticism about critical thinking? About being open to any idea (or none) as long as it survives rational deliberation? Doesn’t this consensus thing mean that the whole movement is actually just kowtowing to white-coated authority? Well, yes and no.

To begin with, let’s remember that there are many people who are strongly skeptical of certain ideas, but who are not counted as part of the skeptical movement. Take Holocaust skeptics, global warming skeptics and evolution skeptics. In the skeptical community, we call them denialists. Why? Because their views go against scientific consensus.

Science presupposes that all participants have a skeptical frame of mind and arrive at conclusions through rational deliberation. If a large group of knowledgeable people working in this way arrive at a consensus opinion, then there is really no good reason for anybody with less knowledge of the subject to question it. Informed consensus is how scientific truth is established. It’s always provisional and open to reevaluation, but as long as there’s informed consensus, then that’s our best knowledge. Humanity’s best knowledge.

Although I see where Martin’s coming from, I found this viewpoint somewhat disturbing, leading me to echo Martin’s own words; Well, yes and no.

For a skeptic, in matters of science it is undoubtedly true that the scientific consensus is always the best place to start an evaluation when evaluating unfamiliar issues. It is certainly possible that the scientific consensus is wrong in almost any area, but the consensus almost always represents the best current understanding of an issue. It is also correct, as Martin argues, that authority matters. Like him, I’m more inclined to accept the pronouncements of someone who has actually dedicated his or her life to studying the issue systematically; i.e., an expert. If the topic is evolution, then that expert would be an evolutionary biologist. If the topic is the Holocaust, then a historian specializing in the Holocaust would represent an appropriate expert. For cancer, an appropriate expert would be an oncologist. The list goes on.

Where I start to have a bit of a problem with Martin’s viewpoint is when I start to contemplate the nature of scientific consensus itself in many areas of science. Not all consensuses are created equal because, depending upon the field, the strength of scientific consensus can vary quite markedly depending upon the topic or even the subtopic within the topic. For example, the scientific consensus supporting the theory of evolution, particularly common descent, is exceedingly strong. It’s one of the strongest of all scientific consensuses. Similarly, the consensus that natural selection is a major driving force behind evolution is very nearly as strong. However, as the discussion devolves into more detailed areas, inevitably the consensus weakens. Eventually, subsidiary areas of a discipline are reached where the consensus is weak or where there is no consensus, such as what the function of “junk DNA” is, whether it is subject to natural selection, and if so how much. (Real evolutionary biologists could probably come up with a better example.) These sorts of questions are often at the edge of scientific knowledge, and it is not always easy to recognize what they are. It is also these issues at the edge of our knowledge that are attacked as proxies for the much more strongly supported core theory. Creationists are notorious for this sort of tactic.

The same is true of many other disciplines, including my own discipline of medicine and surgery. However, the scientific consensuses are rarely quite as strong as the theory of evolution; usually, the strength of the consensus is proportional to the ratio of data supporting it that comes from randomized clinical trials to data from epidemiological studies, the latter of which are more prone to confounding factors. That does not, however, mean that there isn’t a strong consensus about many issues. For example, there is in essence no doubt that HIV causes AIDS, the claims of some notwithstanding. Similarly, there is in essence no doubt that smoking tobacco vastly increases a person’s risk of lung cancer and heart disease, along with a host of other medical problems. Not even the tobacco companies try to argue that anymore. When we come to subsidiary questions, though, the consensus is generally not as strong. For example, it has become increasingly appreciated that secondhand tobacco smoke increases the risk of heart disease and lung cancer in people chronically exposed to it. However, because the effect is smaller than it is for people who actually smoke cigarettes, there is a lot more “noise” in the studies, giving a lot of wiggle room to people who dislike the idea of the government banning smoking to claim that such bans are not scientifically supported, and it’s taken a long time for scientific and clinical studies to firm up the conclusion enough to the point that it is now a strong consensus.

Not surprisingly, given the difficulty doing controlled experiments and the nature of the material, which makes it more easily politicized or influenced by biases, a truly strong consensus is harder to come by in the humanities and social sciences. However, even so, it is not impossible. For example, one of my areas of interest is the Holocaust. There is overwhelming and incontrovertible evidence that the Nazi regime embarked on a systematic program to round up and exterminate the Jews in territories they controlled. Methods used included shooting, hanging, gas chambers, and a system of camps designed to literally work their inmates to death through a combination of grueling labor, insufficient food, and unsanitary and crowded conditions. However, this consensus becomes less clear when various issues surrounding the Holocaust are discussed. For example, there is the whole “intentionalism” versus “functionalism” debate which, in a nutshell, is the question of whether the intent was there from the very beginning of the Nazi regime or even before (a “master plan,” if you will) to exterminate the Jews or whether the Holocaust grew “organically” or “functionally” out of increasing persecution, radicalization of Nazis carrying out the program, and the question of what to do with the millions of Jews that fell under Nazi control after Hitler invaded the Soviet Union. Holocaust deniers love to misuse this debate to claim that Hitler didn’t know and didn’t order the Holocaust.

The reason this whole question came up relates to a debate that had been brewing for a few months at least but recently bubbled to the surface at the 76th Meeting of the Skeptics’ Circle last week over posts submitted by Sandy Szwarc at Junkfood Science. I’ll admit that she’s a difficult problem and one reason why I’ve been thinking about these issues lately as I decide how the Skeptics’ Circle should go forward and what, if any changes, should be made. Most of her non-obesity-related stuff is often actually not too bad. However, whenever she blogs about diet and obesity, there’s usually a problem. And it’s not the sort of thing that necessarily jumps right at you off of her blog. Certainly it didn’t for me. Rather, it’s the sort of thing you have to read her blog closely for a while (which I did) to start to realize. As I read her blog, more and more it bothered me that all of her “skepticism” was inevitably in the direction that being obese is not only not unhealthy but is actually at least as healthy as not being obese, that eating fatty foods is perfectly fine, and that virtually any study she looks at that says that eating fatty foods or too many calories predisposes to health problems is a pile of crap while any pile of crap study claiming otherwise is the latest and greatest. All of this leads her to conclude that virtually every warning made by scientists and physicians about diet is fearmongering. Worse, she has a distressing tendency to use unscientific tactics, such as cherry picking data, attacking consensus, and alleging conspiracies. I also found it telling that, unlike most bloggers, myself included, Szwarc does not permit comments. If there’s one thing that skeptics usually encourage, it’s spirited debate. That’s impossible in a blog that doesn’t permit comments.

The end result of this incident leads me to be a bit uncomfortable with Martin’s blanket statement that “a real skeptic always sides with scientific consensus.” The reason is that what the scientific consensus actually says is not always that clear for many issues, even among those who work in the field. Indeed, there are comparatively few issues in science (evolution, for example) for which a strong consensus exists, and even fewer in the social sciences and history (the Holocaust, for example) for which an equally strong consensus exists. These are very strong consensuses, and to overturn them would require extraordinary evidence, evidence at least equal to the evidence supporting them. Consequently, when someone says that evolution is false or that the Holocaust didn’t happen (or the lessor form of Holocaust denial, that nowhere near 6 million Jews died), it’s fairly easy to recognize such person as a crank and denialist, and one should not hesitate to label them as such.

But what about consensuses that are strong but not as bullet-proof, usually because, although there is a consensus, there are fairly wide error bars around the predictions or uncertainty regarding the importance of various factors? The prototypical example of this is anthropogenic global warming, for which there is a strong consensus among climate scientists but still a fair amount of uncertainty about the outcome. Another example, of course, is the scientific consensus about the link between obesity and adverse health outcomes. How do we differentiate legitimate skepticism about the consensus from denialism?

This is where I tend to agree with Mark Hoofnagle. It’s more about tactics and how evidence is used to support an argument. Scientific skepticism looks at the totality of evidence and evaluates each piece of it for its quality. Cranks are very selective about the data they choose to present, often vastly overselling its quality and vastly exaggerating flaws in current theory, in turn vastly overestimating their own knowledge of a subject and underestimating that of experts. This is perhaps the key characteristic of cranks and the biggest difference between a crank and a true skeptic. In addition, because the mainstream rejects them, there is often a strong sense of being underappreciated, leading them to view their failure to persuade the mainstream of the correctness of their views as being due to conspiracies or money. Antivaccinationists, for example, view the rejection of their belief that mercury in vaccines or even vaccines themselves cause autism by mainstream medicine as evidence that we’re all in the pocket of big pharma. Global warming denialists see the consensus as being politically motivated by the desire of “liberals” to tell them how to live. Evolution deniers view evolution as the result of atheistic scientists wanting to deny God. People like Sandy Szwarc view the consensus that obesity leads to health problems as being due more to moralizing and bigotry against the obese, which, whether it is true or not, is an easy claim to make because there has been and is a lot of bigotry against the obese.

What a lot of this distinction boils down to is that crankery, denialism, pseudoskepticism, or whatever you want to call it tends, either intentionally through ideology or unintentionally through an ignorance of the scientific method, to conflate and/or confuse nonscientific, ideological arguments with scientific arguments. This is not to say that scientists and skeptics are free from their own biases, whether ideological or simply a desired result that they hope to find. Far from it. However, skepticism means applying the scientific method to claims, whatever its faults, scientific method is the best method thus far devised to minimize these biases. As scientists, the reason we use the scientific method is not because we consider ourselves superior to the cranks, but rather because we recognize that we are human too and thus just as prone to falling into the same traps as they. Moreover, we know that science is a work in progress and that what is considered correct today may well be modified tomorrow. This change, however, is not brought about by cranks cherry-picking data but by rather skeptical scientists probing for weak spots in our current understanding, making hypotheses, and then testing whether current theory or the new hypotheses make the better prediction. Thus, being skeptical of the consensus is not the mark of the crank. It’s how and why that skepticism exists that distinguishes crankery from genuine scientific skepticism. We should not forget that.