Skip to content

On making predictions and swallowing the outcome

May 23, 2011

Science is a method of predicting something and then quantifying how wrong you are.  This will vary from “almost right” to “not even close”, but few hypotheses survive experiments intact.  Although scientific progress means that each prediction made about a specific issue should be increasingly accurate, one must always be prepared, at least in principle, to completely abandon an idea or theory when compelling evidence is presented to do so.  This week, an evangelical radio broadcaster and a significant number of other Americans predicted that the world would end at 6 pm on Saturday 21 May.  Since it is currently Sunday 22 May, we can logically conclude that this prediction was complete baloney.  In fairness, this prediction was not based on any rational understanding of anything at all, which largely explains its failure (and “rapture” has itself been predicted repeatedly in the past).  Newton, it is claimed, predicted 2060 as the true date, so expect the media to go nuts for this nonsense again at some point – after all, Newton invented gravity…

When I was a PhD student, I promised myself I would avoid, as far as possible, building a comprehensive personal model of how the body worked.  My initial experience with science at the “coal-face” led me to conclude that because so much of what seems to be taken for granted is either nonsense, or possibly a bit wrong, it would be futile to build a model based on anything other than direct experimental work.  All else should be considered provisional.  This was reinforced by my PhD examiner, who stepped in to spare my stumbling over a particularly tough question.  I’d spent 10 min talking it through before he demanded a straight answer.  “I don’t know”, I said. “If you know something, say it definitely,” he replied, “and if you don’t know something, SAY IT DEFINITELY!”.

Scientists, by and large, are comfortable with admitting they don’t know things.  Many other people are not.  Politicians, for example, seem to spend much time preparing themselves to avoid “don’t know” at the dispatch box, on TV and radio.  Those of religious faith also have difficulty with uncertainty.  This perhaps best illustrated by those involved in the “rapture” movement who claimed that the lack of any outcome on Saturday was God’s way of testing their faith.  People often use the same excuse when confronted with tragedy that must, by definition, have been preordained.  When religious predictions fail, those who believed them will never admit error.  This position is easily dismissed.

There is a second group who will never acknowledge uncertainty but are far more dangerous.  These people either fabricate or cherry pick evidence so as to support a conclusion they already hold and will never let go.  Perhaps the most dangerous group in this regard are climate change deniers.  Last week, I made the error of visiting James Delingpole’s blog.  As usual, he was trying to discredit the idea that man-made global warming was happening, and he did so with one of the most outrageous examples of cherry picking I have ever seen.  He had presented a figure showing what was purported to be the global air temperature record from 1880 to 2000, with subsequent predictions based on the IPCC report and a trend based upon a multi-decadal oscillation.  The graph itself came from a Global Warming Policy Foundation report written by Lord Turnbull, and he got the graph from a paper written by Akasofu (2010) in Natural Science – a journal in its third year of publication that seems to publish quite a lot of papers sceptical of climate change.

The graph presented in the paper of Akasofu (Figure 9) cherry picks in a number of respects.  For example, it uses the CRU record from 1880 to 2000, in spite of the fact that their records extend considerably further (1850-2010).  The graph assumes that 1. global temperature is recovering in a linear fashion from a “little ice age”, 2. there is a multi-decadal oscillation superimposed on this overall trend, and 3. this trend will continue ad infinitum.  This scenario is presented in contrast to the IPCC estimates of temperature change to 2100.  The IPCC prediction, naturally, quantifies the error in the estimated temperature, the Akasofu trend, revealing, does not.  Finally, to validate the linear trend with decadal oscillation, the CRU record of global air temperature for 2008 is plotted and is shown sitting almost perfectly on the line predicted by Akasofu.  You’d be forgiven for wondering why 2008 was used in this analysis.  Therein lies the final example of cherry picking: 2008 happens to be the lowest annual temperature measured by CRU this century; as you can see from the link below, had 2007, or 2009, or any year in the 21st century apart from 2008 been plotted, it would have agreed with the IPCC scenario and not that of Akasofu/Turnbull/Delingpole:

CRU’s temperature record 1850-2010

The GWPF have form with respect to being selective about temperature records.  It is from them that the meme that global temperatures have been declining since 1998 has been spread in the UK, a meme totally destroyed by George Monbiot in an exchange with Ian Plimer.  Cherry picking of this kind serves only one purpose: to provide evidence that supports a view held before the analysis was done, rather than a conclusion reached by a dispassionate analysis of ALL the available evidence.  Had that been done in this case, there would be no basis for the “validation” performed, nor the linearity continuing backward from 1880.  Thus, James Delingpole, who admits he has never read a scientific paper on any subject, and considers himself “an interpreter of interpretations”, parrots the erroneous citation of an erroneous analysis presented in a journal sceptical of climate change.  Overall, this is a third-hand interpretation of total nonsense.

It is rather amazing that Delingpole and his ilk can present these interpretations without being ashamed of them.  They almost certainly consider the data damning, but, to paraphrase my external examiner once more “being in possession of a red flag and a bull does not make one a matador”.  Interpretations lie, data don’t.  That’s why presentation of all the available evidence is a sacred duty.  Any less is the path to bullshit.

One Comment leave one →
  1. June 18, 2013 7:45 pm

    o nie posiada? A także wszystkim abandonner
    *Merle* to
    replikuje? Pasterzom co więcej, tymże odkąd owieczek?

    Krzyżak rozejrzał się po raz drugi, choć pobliscy goście,
    spoczywający w malowniczych pozach na ławach nie sprawial.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: