Monday 23 November 2009

Hanson on HadCRUT [updated]

Hanson's reaction to the computer hacking release of a bunch of emails from the climate researchers is the same as mine:
Joel Achenbach comments:
This is not a scandal so much as a window on real scientists working on a politicized issue. … “Gravity isn’t a useful theory because Newton was a nice person.” I agree. But isn’t it also true that Newtons antipathy towards Hooke and his use of his position in control of the Royal Society, ensured that the concept of an achromatic lens for a telescope … had to wait until after [Newton's] death.
Yup, this behavior has long been typical when academics form competing groups, whether the public hears about such groups or not. If you knew how academia worked, this news would not surprise you nor change your opinions on global warming. I’ve never done this stuff, and I’d like to think I wouldn’t, but that is cheap talk since I haven’t had the opportunity. This works as a “scandal” only because of academia’s overly idealistic public image.

It is a shame that academia works this way, and an academia where this stuff didn’t happen would probably be more accurate. But even our flawed academic consensus is usually more accurate than its contrarians, and it is hard to find reliable cheap indicators saying when contrarians are more likely to be right.
On the plus side, it's good that public skepticism will be raised a bit. On the other, I'd worry that folks will only use such skepticism against academic results that conflict with their priors and be too dismissive of correct consensus that leans against them.

Update: I'm not trying to downplay too much the stuff that's in the emails. It looks like the consensus on warming was weaker than was publicly portrayed. It may be the case that data was massaged excessively, but that would be a conclusion that could only really be drawn after somebody familiar with the field had a check through things. You could probably find similar emails among macro econ researchers saying things like "well, nothing made sense until I ran the data through an HP filter to clean it".

Update 2: Tyler Cowen's take is similar to Robin's as well.

6 comments:

  1. Was that last line an attack on macro or a defense of warmists?

    I only remember one thing from my honours macro class: half way through the year the lecturer mentioned, in a disturbingly off-handed way given the pain he was inflicting, that "by the way, the jury is still out on whether macro has any empirical validity whatsoever". At the time IIRC he thought another forty quarters of data would settle the matter. Is the jury in yet?

    ReplyDelete
  2. "You could probably find similar emails among macro econ researchers . . . "

    Given the abject failure of macro econ to work out what the hell is going on the world -- not least, not to have even noticed when the economic world was beginning to collapse all around them -- I wouldn't be at all surprised.

    ReplyDelete
  3. @Matt: ambiguity can be fun.

    I'd ignore entirely the bits where the warmists talk about how much they hate the deniers. You'd have identical emails on both sides, I'd expect.

    Some data-cleaning techniques are entirely legitimate and even necessary to be able to draw out trends: have to filter out some of the noise and figure out ways of aggregating data from diverse weather stations around the world where the number of observing stations is increasing over time and use that to adjust the prior trend from when there were fewer stations.

    We always come to data with models, even if they're just implicit. That's one reason why so many economists discounted the Card and Krueger minimum wage study out of hand. Sometimes observations sufficiently out from a model can point more to a problem in the data or the data analysis than in the model.

    I really really hope that the result of all this is that Hadley releases its raw data for others to look through. It's hard not to see a recent warming trend, albeit now plateaued, in that data. But if the released data is a sanitized distorted mess, well, that's something different.

    ReplyDelete
  4. You need to examine the HARRY_READ_ME.txt file and ask yourself if these data management practices are acceptable for an economic and political agenda item of such importance?

    ReplyDelete
  5. Quick flip through -- sounds like a new guy trying to figure out a prior coder's very opaque code. Messy!

    ReplyDelete
  6. @mrv: you were entirely right on HARRY_READ_ME.

    ReplyDelete