Friday, December 4, 2009

Penny Wise on Climategate

The most sensible comment I have seen on the whole climate-change email scandal now known as Climategate was by Clive Crook of the Atlantic:
Climate scientists lean very heavily on statistical methods, but they are not necessarily statisticians. Some of the correspondents in these emails appear to be out of their depth. This would explain their anxiety about having statisticians, rather than their climate-science buddies, crawl over their work.
That hadn’t occurred to me but is obviously true. Science and statistics may be bedfellows but they are not at all the same thing, and expertise in one area does not confer expertise in the other.

So, as I do when the maths gets too hard, I asked my economist friend Penny Wise for a comment. She writes:
Yes, it is not hard to see how and why a group of scientists decide to just do the statistics themselves, all having done a little bit of statistics in their science courses. But statistics (or in economics, econometrics) is a specialty like all the rest and, having struggled through a second-level econometrics course myself, I know how hard it is.

It is also very easy to do badly, as you can use any number of off-the-shelf stats packages to do the work for you in a rudimentary way. I recall an hilarious paper where some top-level econometrician absolutely shredded a paper by Milton Friedman.

Macroeconomic models of the economy are unstable enough and require all sorts of “tricks” and judgements to keep them giving sensible results. Portfolio optimisation models are wildly unstable. So it is no surprise that a full climate-change model, which must be massively complex (in the sense of having a huge number of correlated variables) must reflect a large number of scientists’ judgements.

The good to come out of this will, I hope, be to force data and methodologies to be made widely available to scrutiny. As I understand it, you cannot publish in a quality econometrics journal unless you also provide your data and methods. If nobody else can replicate those results, there is a problem.

It is astonishing that NIWA has been so protective of this for the NZ temperature record. I don’t doubt they have done a reasonable job of it, so why not put it out for review?

No comments: