So-called “climate-gate” involved a number of accusations that concerns about global warming are based on bad and fallacious reasoning. As the deeper analysis comes in, the email messages from the University of East Anglia turn out to be rather unremarkable if a little protective and no cause to believe that the reasoning behind climate concerns is poor. However, the initial accusations of “bad science” have given authority to people whose reasoning about the issues is itself manifestly poor, as evidenced by today’s appointment of University of Western Ontario professor Chris Essex as the official climate advisor to the President of the World Federation of Scientists.
Essex’s reasoning on this subject is notably bad. Not only does he maintain contrary to the considered evidence that the UEA researchers behaved irresponsibly, but he also publicly and repeatedly employs bad analogies to defend his own climate scepticism. Essex argues that temperatures are like phone numbers, lacking relevance to their means. From The National Post:
“Many people think that you can make sense out of an average of anything at all. My usual reply is to ask what an average over telephone numbers means. Temperature is like that. When averaged, it does not produce an actual temperature of anything, any more than an average over telephone numbers must be a callable number, let alone a number you might care to call.”
That analogy neglects the manifest empirical relationship between a temperature reading and a climactic situation, compared with the randomness of whether a phone number can be called. Sure Essex claims to have stumped a statistician with his analogy, but can he actually stump a climate scientist?
I suspect, or at least hope, he couldn’t stump my argumentation students. I’m quite sure he couldn’t stump my social epistemology students, who have learned a lot about expertise this term. Mathematicians and physicists seem to have special desire to make general pronouncements about other fields in which they lack expertise. This is much like the credentials of the “scientists” on creationist websites who are not actually biologists but in abstract fields.
(Thanks to Wayne Myrvold for pointing out this appointment and how terrible it is.)
RAIL 
If Essex claims to have stumped a statistician with that claim, he was talking to a spectacularly bad statistician.
Temperature is an interval scale, so the mean makes sense as a statistic. See: http://en.wikipedia.org/wiki/Level_of_measurement
But telephone numbers aren’t that kind of thing at all. This seems like something every statistician (every statistically literate person, even) would know.
Also, the link to the National Post doesn’t work.
Thanks Seamus. I’m glad to understand more about the details, and the link to The Post should work now.