Monday, February 21, 2011

Debunking Science: Why It Can’t Be Done

I will never forget the shock I experienced last fall when I saw a respected nutrition researcher attempt to explain away previous studies that clashed with many of his findings. That scenario, with scientific studies pointing in opposite directions and researchers disagreeing, is to be expected, but the researcher was not merely trying to point out the limitations or possible errors in the earlier studies. Rather, he was trying to discredit the entire field of epidemiology, in which the studies were done.

Without going into the whole story, this was a substantive scientific conflict. Clinical observations suggested that certain nutrients cured certain diseases in about half of cases, a highly impressive result. However, all epidemiological evidence says that foods containing the nutrients cause more illness than they cure and result in people dying sooner. If both are true, this merely means that for the average person, the treatment is worse than the disease. That would hardly be a surprise — the main reason we don’t apply medical treatments to people of average health is that the intervention is likely to make them sicker — but it is nevertheless a conclusion that can be debated. One would hope, though, that the debate would go on in a scientific way, that it would not immediately degenerate into a scientist in one field trying to debunk an entire field of science that he happens to disagree with at the moment.

Alas, when the debunkers come out, it is usually science that is their intended victim. This is confusing because most debunkers present themselves as scientists or friends of science, yet scientific experiments and demonstrations are not their tools of choice. Rather, they argue. The essential argument of a debunker is something along these lines: “The events these people claim to have observed could not have occurred because of the principles of science.” They use scientific dogma to explain away experimental results and real-world observations.

If you are familiar with the process of science, you quickly realize that if debunkers have their say, there won’t be any science in the end. Scientific principles are supposed to be created and supported with observation and experimental evidence. If experimental evidence can be selectively explained away, then anything at all could be established as a scientific principle. If we ever get to that point, with certain points of view held to be infallible, science will have turned into a religion. Thus, debunking and science are generally at odds with each other.

Strangely enough, the winter season currently winding down in North America has also been a debunking target. “There is nothing unusual about this winter’s weather,” the debunkers tell us. “It’s winter. It’s supposed to be cold and snowy.” Notice the familiar pattern of debunking. There is an appeal to science, in this case, the climatological profile of North American winters, in order to discredit actual observations.

The attempt to debunk the winter of 2010–2011 doesn’t stand up even when you look at the days of weather observations one by one and note the many records set: the coldest temperature ever recorded in some places, the deepest snowfall of any single storm in others. But the distinct nature of the winter season jumps out at you when you look at the season as a whole using statistics. These are statistics anyone can compute, such as the total snowfall for the season (by location, of course), the average temperature for the months of December and January, and the number of “epic” U.S. winter storms (count them: 3). Those who would try to debunk this winter’s weather are left in the uncomfortable position of trying to debunk the field of statistics.

That’s an obstacle that would stop any sensible person, but there is nothing very sensible about debunking in the first place. Debunkers, after all, are the people who go to talk to people who have seen something extraordinary and explain to them that what they observed could not have taken place. It’s not something you can do if you believe in obstacles. Therefore, it’s not so surprising if you see a debunker take on, and try to take down, the idea of statistics.

This is happening, for example, among the bedraggled supporters of the global climate stability hypothesis, which essentially says that climates don’t change, at least not by very much, and not for very long. It is fair to describe this group as climate change debunkers, as there really is no observational support for the idea that climate cannot change. We see the weather change every day, and historical records are clear enough about weather patterns changing from one century to the next. If you don’t believe in records, you can go to places where the permafrost is melting and directly observe climate change taking place. But the idea of global climate stability hangs on, supported by a sort of “common sense” notion that Earth is too large to have the kind of change that would show up as climate change. As always in debunking, there is a scientific principle to support this idea. We know from experience that the larger something is, the slower it moves. It takes more people pushing to move a heavier car, a longer time to boil a larger pot of water. The experience of homeostasis, another idea verified by science, also supports the idea of climate stability.

But this only means that climate stability is a reasonable hypothesis, not that it is true. With observation telling us that climates are stable only in select places and for limited periods of time, supporters of the global climate stability hypothesis have decided to try to debunk the statistic of global average surface temperature. Unfortunately for them, they can’t undo this statistic by questioning its assumptions. Global average surface temperature is a highly robust measure, resulting in much the same picture no matter how you mess with the assumptions and methodology along the way. The only line of argument left is to say that statistics in general aren’t particularly meaningful, and these days, it is easy to find adherents of the global climate stability hypothesis saying just that.

This is what all debunking comes to in the end. When discounting individual observations is not enough, you have to start throwing out pieces of the scientific method one by one until you are taking a stand against all science. It’s worth noting than many of the most famous debunkers have given it up after a number of years. There is one who continues his stage performances, but now uses them to teach his audiences some of the phenomena of perceptual psychology. It is almost a complete turnaround, from “There is nothing to see” to “Here is how to observe carefully.” Without knowing any of the individual stories, I can only guess what might happen to make a person turn from debunking to science. There is a predicament that every debunker will inevitably face, sooner or later: a group of people with a story so credible or a demonstration of a phenomenon so inexplicable that the debunker cannot even convince himself that nothing happened. These people clearly experienced (or demonstrated) something, but what? Then, or separately, debunkers may start to own the scientific garb that they present themselves in, or they may be taken with natural human curiosity, and start to ask scientific questions: “What is really going on? How can it be proved or demonstrated?” Once you start asking questions, of course, you are making the move from debunking back into science.