Sunday, July 16, 2006

Daily Policy Digest | NCPA

Daily Policy Digest | NCPA:

HOCKEY STICK HOKUM

Before 1999 the accepted view on climate change was that the world had undergone a warming period in the middle ages, followed by a mid-millennium cold spell and a subsequent warming period -- the current one.

That all changed when paleoclimatologist Michael Mann's research paper eliminated the Medieval warm period from the history books. With a nice, steady temperature oscillation that persists for centuries followed by a dramatic climb over the past century, Mann's work produced the "hockey stick" graph.

The trouble is that there's no reason to believe Mann, or his "hockey stick" graph of global temperature changes. Subsequent studies have shown Mann's analysis to be less than definitive:

  • In 2003, Ross McKitrick and Steven McIntyre published an article in a peer-reviewed journal showing that Mann's methodology could produce hockey sticks from even random, trendless data.
  • Furthermore, in a soon to be released report by the House Energy and Commerce Committee, the three researchers -- Edward J. Wegman of George Mason University, David W. Scott of Rice University and Yasmin H. Said of Johns Hopkins University -- find that Mann's methodology is biased toward producing "hockey stick" shaped graphs.

In addition to debunking the hockey stick, Wegman goes a step further in his report, attempting to answer why Mann's mistakes were not exposed by his fellow climatologists. His conclusion is that the coterie of most frequently published climatologists is so insular and close-knit that no effective independent review of the work of Mann is likely.



Now, the National Center for policy Analysis is hardly known for having no position on these things, but I'm in the process of reading the Wegman report, and I don't think this is massively overstated.

I'll be writing more about the report later today.

3 comments:

Rick Ballard said...

I think that Mann is just as great a scientist as Lysenko ever was. His ability to construct a believeable hypothesis may, in fact, be greater than Lysenko's. It's the scientific world's credulity that is on trial here - and it appears to be failing the rather simple test of "Are you going to believe me or your lying eyes".

chuck said...

His conclusion is that the coterie of most frequently published climatologists is so insular and close-knit that no effective independent review of the work of Mann is likely.

Yep, these little coteries do develop. I recall reading papers on the determination of molecular potentials using the Rydberg-Klein-Rees theory. There was one whole set that referenced each other, then a later set centered around Richard Zare's work. Apparently Zare wasn't aware of the earlier work, or he simply ignored it, because I don't recall him citing it. A quick google shows that Zare's students at Stanford continue to publish work in the field. Anyway, I thought at the time that the insulation of the two groups from each other was quite astonishing.

Heh, and I note that the parametric methods I spent some time on are now coming into use. Looks like I might have been the first in a few things, I just didn't publish.

Charlie Martin said...

Anyway, I thought at the time that the insulation of the two groups from each other was quite astonishing.

I had much the same experience with formal methods: there was the Oxford group, and the American group. If you did a transitive closure of the Oxford groups citations, you'd never get out of the Oxford group.