See Comments down arrow

A statistician recants

01 Nov 2023 | Science Notes

In 2015 the BBC ran a program called “Climate Change by the Numbers”, in which they picked three numbers they thought were representative of the climate change issue and handed them to three professors previously uninvolved with that topic and asked them to spend half an hour explaining what they meant. The numbers were: the average warming since 1880 (0.85C), the degree of certainty that at least half the warming since 1950 is man-made (95%) and the cumulative amount of carbon that can be emitted if warming is to remain below “dangerous levels” (one trillion tonnes). The 95% certainty number was handed to Dr. Norman Fenton, a now-retired Professor of Risk at Queen Mary University and author of over 350 peer reviewed articles in probability and statistics. He spent his half hour not discussing climate change, which he was not familiar with, but how models are constructed for other kinds of analysis like football scores, which he happened to be working on at the time. He said it was an analogy for how climate science worked, only there wasn’t enough time to look at the topic in detail, but he assumed that somebody somewhere has shown that scientists are 95% certain etc. etc. Some time later, however, he finally had the time to study it, and in a new report for the Global Warming Policy Foundation he says he got it wrong, the IPCC made a mistake and the 95% certainty claim is baseless. Wanna bet the BBC won’t be asking him back?

The problem, explains Professor Fenton, is one that is sufficiently familiar to have a name, “The Prosecutor’s Fallacy”. It happens when people reverse the condition and the conclusion, and is also called the “fallacy of the transposed conditional” by professional philosophers trying to repel their audience with the same skill sometimes displayed by statisticians. But here’s Fenton’s plain-language illustration:

“If an animal is a cat, there is a very high probability that it has four legs. However, if an animal has four legs, we cannot conclude that it is a cat. It’s a classic error, and is precisely what the IPCC has done.”

In this case if warming is man-made, there is a very high probability that it is unusual. But just because it is unusual, we cannot conclude that it is man-made.

The logic in the IPCC report begins with the claim that, according to models, most warming is man-made. And if that proposition is true then, if there was no man-made impact, there would be less than a 5% chance of seeing the warming that has happened since 1950. But since we see the warming, it means there is less than 5% chance there was no man-made impact, hence a 95% chance most of the warming was man-made.

Wrong, says Fenton. Backwards, in fact:

“The problem is that, even if the models were accurate (and it is unlikely that they are) we cannot conclude that there is at least a 95% chance that more than half the warming was man-made, because doing so is the fallacy of the transposed conditional. All we can conclude is that there is at least a 95% probability we would not observe the warming we have seen based on the climate change model simulations and their multiple assumptions.”

But we have to take account of the possibility the models are wrong, he says, or that something other than greenhouse gases can cause warming. And just because there’s only a one-in-twenty chance of something happening, it doesn’t mean if it did happen it had to be man-made. It just means it was unlikely. But unlikely things happen quite often; given how many things happen, it is highly unlikely that a lot of them wouldn’t be unlikely.

As Fenton adds, however unlikely warming may be, it has happened frequently in the past. And of course:

“previous periods of warming certainly could not have been caused by increased greenhouse gases from humans, so it seems reasonable to assume – before we have considered any of the evidence – that the probability humans caused most of the recent increase in temperature to be very low; only the assumptions of the simulation models are allowed, and other explanations are absent. In both of these circumstances, classical statistics can then be used to deceive you into presenting an illusion of confidence when it is not justified.”

The IPCC’s conclusion is just circular reasoning. All warming is man-made, we see warming, therefore it’s man-made. Unless it isn’t.

If Professor Fenton keeps digging he will find that the IPCC has already dealt with this objection about past warming by wiping out past climate changes and making it look like temperatures never did anything until Henry Ford invented the Model T. Although he seems to be aware of that issue too because he makes an offhand reference to the hockey stick and the statistical problems that have been identified with it.

It’s too bad that it took eight years for Professor Fenton to realize his error, but better late than never, and kudos to him for coming forward about it. It may be too late to save the UK from freezing this winter with their heat pumps and their iced-over solar panels but at least it’s a step in the direction towards sanity.

4 comments on “A statistician recants”

  1. Fenton never busted anything, Some of his objections were unfounded and based on a lack of knowledge of anatomy and how Viruses work, What he attacked was the ONS data set not the efficacy and need for vaccinations in preventing serious illness because of being infected with Sars-Ncov2,

  2. Gregory, Sars-Cov-2 is about 10,000 times as dangerous to the old and infirm compared to the young and healthy. To think that the same treatment should be given to the entire population was always obviously idiotic.

  3. It's nice to know that at least someone out there
    Is willing to admit that they could be wrong.
    I find it interesting, that in an older IPCC report, that John Kerry and AL Gore liked using, clearly showed that a warning trend started BEFORE an increase in CO2 gas by 30 Years or more. By the IPCCs own report, the increase in CO2 is caused by a warning trend NOT a warning trend caused by an increase in CO2.
    Let us Not forget the effects water vapor has on a warning trend! Or is that too much of an "inconvenient truth "??

Leave a Reply

Your email address will not be published. Required fields are marked *