Wednesday, October 20, 2004

Global Warming and Internet Ping-Pong

Yesterday i saw a pointer in Andrew Sullivan's blog to an article at Technology Review about evidence that the now-famous hockey stick graph that appears to show a consistent warming trend during the 20th century was produced through dubious statistical means (basically, an incorrect usage of principal component analysis). The Technology Review article, written by McArthur fellow Richard Muller, discusses information published by a couple of guys named McIntyre and McKitrick. Muller apparently considers their results to be definitive evidence that the the hockeystick diagram is bogus. However, the CrookedTimber blog has a discussion of analysis done by an Australian computer scientist Tim Lambert on a subsequent paper by McKitrick and Pat Michaels that suggest McKitrick maybe sort-of doesn't know what the hell he's talking about.

So after bouncing around between several different sites, most of which seem fairly credible, i not only don't know what the truth is, i'm probably more confused than when i started. The only option left would be to go to the original sources and try to understand the analyses performed and then read the various accounts of the mistakes made, etc. I've made two completely unrelated conclusions from this little odyssey. First, the Internet+Google might equal something amazing, but we've got a long way to go before it can be used to support decisions. Second, i conclude that global warming opponents will be able to use this confusion to much greater effect than environmentalists.

No comments: