Saturday, 20 February 2010

How Skeptics Distorted the CRU Emails In the Name of "Climategate" (Part 2)


I continue my critical analysis of the SPPI "Climategate Analysis" report, email by email. I will try to stick to the following formatting to make things clear:
Parts in this color and in bold are the reports own words. Parts in this color and in italics are the report's quotes from the CRU emails. All other parts are my own analysis.

email 0843161829

The report starts:

Two days after the previous exchange, Gary Funkhouser reports on his attempts to obtain anything from the data that could be used to sell the message of climate change:

I really wish I could be more positive about the ... material, but I swear I pulled every trick out of my sleeve trying to milk something out of that. ... I don’t think it’d be productive to try and juggle the chronology statistics any more than I already have—they just are what they are ... I think I’ll have to look for an option where I can let this little story go as it is.

Now where in that text is there anything about "selling the message of climate change"? That interpretation cannot have been drawn from the email, which simply doesn't support it. So where does it come from? Where is the hidden place that the Analysis frequently dips into to draw these wild accusations?

What are Briffa and Funkhouser even discussing here? The self-styled "Climategate Analysis" doesn't say. That's right the skeptics can't even be bothered to figure out what aspect of the science is being discussed. In fact the report goes as far as to omit parts of it with ellipses as if it's all just getting in the way of the allegations they want to make.

Here's a thought - what if Gary Funkhouser is not complaining that he can't obtain data to sell the message of climate change, but is actually complaining that he can't get data to show anything? Ie that the data is too noisy to yield a signal? Wouldn't that be consistent with the email? Why yes it would. In fact one of the ellipsed out sentences "The data's tempting but there's too much variation even within stands." suggests that no?

Me thinks you "skeptics" have the idea of email analysis backwards! You are interpreting the emails to fit your conclusion. But what you should be doing is interpreting the emails to determine the conclusion. In other words the "hidden place" mentioned earlier which you are drawing to interpret individual emails is in fact your conclusion.

So I see a vicious cycle, a kind of snake eating it's own tail. Your conclusion comes from your interpretation, but at the same time your interpretation is drawn from your conclusion. So which came first, the chicken or the egg? I suspect the paradox is resolved by observing the baggage skeptics arrived with before they began the analysis. They wanted climategate so badly that they formed their conclusion before looking at the emails, which due to their bad analysis technique tainted their interpretation of the emails, which then looped back to reinforce their conclusion. There is surely not enough ridicule in the whole world that would be sufficient to heap upon these "skeptics".

The report continues:

His reluctance to report a “null result” (namely, that the data do not show anything significant) is extremely disturbing as it flies in the face of standard scientific practice, which requires that all results be reported.

Bullshit. If you can't extract a signal from some noisy data and therefore you have nothing to use for any analysis, you don't go and publish a paper saying you couldn't do any analysis because the data was too noisy. Where are these "we didn't do shit" papers published? In the International Journal of Having a Laugh?

The fundamental problem is that any censoring of results that do not lead to a predetermined conclusion will always—by design—bias the corpus of reported results towards that conclusion

Wow that describes the SPPI report. Censoring of the straight-forward interpretations of emails that do not lead to the predetermined conclusion, "bias the corpus of reported results towards that conclusion". However in case of this email, there is no hint of a predetermined conclusion being discussed. Indeed the email seems to be saying the problem is that no conclusion can be drawn.

in the same way that a gambler who always brags about his wins (but stays silent about his losses) will appear to be hugely successful, even if his losses have, in reality, far outweighed his winnings (as is generally the case, in the long run, except for the extremely skillful).

They were clever to add the "except for the extremely skillful" caveat at the end, if they hadn't done it someone might have criticized them over such an error!

We will, sadly, see that this fundamental scientific flaw—which, in and of itself, is sufficient to render the evidence for climate change completely unreliable and scientifically worthless—is one that runs throughout the entire Climategate saga.

They make the word "sadly" sound so...insincere. It's like they are trying to play on my emotions...but utterly failing. The hyperbole here based on the flawed analysis pointed out above needs no comment. We will, sadly, see a lot more of this in the rest of the analysis/saga I am sure.

Note, also, the immense power wielded—albeit ever so subtly—by Briffa: he influenced the analysis that Funkhouser performed, simply by telling him that the results would need to politically “saleable”. Scientists are not naive: they know that securing funding, publication of their papers, and interest from other institutions are the key factors determining their future.

This is talking about the last email I discussed and it showed no such thing. This is a new phenomenon we have discovered though - one email's misinterpretation being relied on as a crutch to misinterpret another. I imagine that quite a lot of the analysis works like this, a rotten construction.

No comments:

Post a Comment