RealClimate.org this past week linked an older Real Climate post titled “Peer Review: A Necessary But Not Sufficient Condition“. The post makes the point that peer review is a necessary condition but not a sufficient condition for credibility, observing that bad papers sometimes make it through peer review. (The bad papers they want us to ignore are all skeptical of a human impact on climate change.)
In fact, I would say peer review is neither necessary nor sufficient for credibility. Bad and even horrible papers are sometimes accepted. Good papers are sometimes rejected. Some good papers, for one reason or another, are never submitted to peer review. So peer review is not a magic wand; it is simply a process that adds value by subjecting scientific work to skeptical scrutiny.
Anyway, one of the post’s main examples of the failure of peer review is a paper by Ross McKitrick and Patrick Michaels (a prominent global warming skeptic) that purported to find economic signals in the temperature records. (It doesn’t really matter what this means. The point is that the non-skeptics didn’t like it.)
It turns out that the work was flawed, because McKitrick and Michaels’s data was in degrees but their trigonometric library measured angles in radians. They acknowledge the error. (ASIDE: They also found that the correction does not affect the overall result. On the other hand, Real Climate alleges that there are other problems with the paper as well.)
But here’s the point: The error was discovered because McKitrick and Michaels made their code available! If they had withheld their code, no one ever could have found the error.
Alas, withholding the code seems to be a common practice in the climate science community. The Hadley CRU would not release its code, and it took a leak to expose the fact that its code was broken. (And unlike the previous case, CRU’s code does not admit an easy fix, or perhaps even any fix at all.)
In climate science, the facts may be on the majority’s side (although I’m not as confident as I once was), but they can learn from the skeptics something about the process of science.
Internet Scofflaw 