"Nothing in it is correct"

The eminent statistician (and occasional BH reader) Radford Neal has been writing a series of posts on global temperature data at his blog. There are three so far:

What can global temperature data tell us?

Has there been a pause in global warming?

and finally

Critique of "Debunking the climate hiatus", by Rajaratnam, Romano, Tsiang and Diffenbaugh.

They are all rather technical but very well written - the clarity of thought is striking. But I particularly recommend the last one, a gloriously deadpan take on a much-trumpeted paper (one which trashes claims of a hiatus, apparently), with gems like this:

The authors are all at Stanford University, one of the world’s most prestigious academic institutions. Rajaratnam is an Assistant Professor of Statistics and of Environmental Earth System Science. Romano is a Professor of Statistics and of Economics.Diffenbaugh is an Associate Professor of Earth System Science. Tsiang is a PhD student. Climatic Change appears to be a reputable refereed journal, which is published by Springer, and which is cited in the latest IPCC report. The paper was touted in popular accounts as showing that the whole hiatus thing was mistaken — for instance, by Stanford University itself.

You might therefore be surprised that, as I will discuss below, this paper is completely wrong. Nothing in it is correct. It fails in every imaginable respect.

...and this:

Rajaratnam, et al. describe [their] data as “the NASA-GISS global mean land-ocean temperature index”, which is a commonly used data set, discussed in my first post in this series. However, the data plotted above, and which they use, is not actually the GISS land-ocean temperature data set. It is the GISS land-only data set, which is less widely used, since as GISS says, it “overestimates trends, since it disregards most of the dampening effects of the oceans”.  They appear to have mistakenly downloaded the wrong data set, and not noticed that the vertical scale on their plot doesn’t match plots in other papers showing the GISS land-ocean temperature anomalies.

You have to read it now, don't you?

Comments

[Richard Tol]

@Ken
Economics does not suffer political conformism.

Your political convictions are fine with me. Your physics-reductionism is bemusing in someone of your age and seniority, but as a friend of mine used to say "blessed are the naive". I object, however, to your unwillingness to discuss statistical issues on their statistical merits, and to your repeated apologizing for sloppy research.

[Richard Tol]

Returning to the topic of the thread, Radford Neal is a pretty solid statistician: Google Scholar puts him in the global top 20 of Bayesian statisticians. His blogposts and his response to the comments show that he has taken time to study the issue.

I quite like his approach of looking at the data before reading the literature. His are a fresh pair of eyes, untainted by the now very large body of literature on the statistical properties of climate data.

[...and Then There's Physics]

Richard,

your repeated apologizing for sloppy research.

Since I've never done so, this is simply untrue. Continuing to assert this, does not make it true.

I object, however, to your unwillingness to discuss statistical issues on their statistical merits

I don't hugely care what you object to. I for example, particularly dislike that you regularly say things that you must know are not true, but I would never judge your department or your university because of that. Doing so, in my view, would be juvenile.

I quite like his approach of looking at the data before reading the literature. His are a fresh pair of eyes, untainted by the now very large body of literature on the statistical properties of climate data.

I think is a particularly silly thing to do and quite likely to lead to all sorts of claims that turn out to wrong because he will not have familiarised himself properly with the basics of the topic before looking at the data.

[Richard Tol]

@Ken
Sorry for being unclear.

Radford has familiarized himself with the basic geoscience and with the sampling method, but not with the existing statistical literature on the subject. This maximises the risk of reinvented wheels but minimises the risk of repeated blind spots.

[Slywolfe]

"It is the GISS land-only data set, which is less widely used, since as GISS says, it “overestimates trends, since it disregards most of the dampening effects of the oceans”."

Did the GISS author(s) really mean "dampening," or should this have been "damping?"

[JamesG]

There is a good article in Nature mag this month about astrophysicists being very prone to the belief that a good theory should override any data that contradicts it; particularly with respect to string theory and the willingness to invent other dimensions to solve a paradox. Apparently there is quite a stramash going on between the theorists and empiricists in that field. In other words 'And Then There's MY Physics'.

[golf charlie]

aTTP 11:58 you are very quick to criticise and accuse.

But why not Mann, Lewandowsky and Cook et al?

[stewgreen]

Good one @JamesG ..ATTMP 'And Then There's MY Physics'.

or ATT-HD ( 'And Then There's His Dogma).. wonder if Ritalin works for it ?

That Nature article

..the accusation that branches of theoretical physics have become detached from the realities of experimental science.
At stake is the integrity of the scientific method,

“Faced with difficulties in applying fundamental theories to the observed Universe,” they wrote, some scientists argue that “if a theory is sufficiently elegant and explanatory, it need not be tested experimentally”.

Some argue that string theory is worth pursuing whether or not experiments will ever be able to measure its effects, simply because it seems to be the ‘right’ solution to many quandaries.

... physicists have begun to use purely theoretical factors, such as the internal consistency of a theory or the absence of credible alternatives, to update estimates, instead of basing those revisions on actual data.

[Pointman]

Posterity is a very nice girl. She'd have nothing to do with the likes of ATTP.

Pointman

[golf charlie]

Pointman & stewgreen, why is aTTP so quick to criticise threats to him, but not threats to the integrity of science, such as Mann, Lewandowsky, Cook et al, M O'Brien and other known associates of his?

[Gras Albert]

and Then There's Politics mindset can be summarised thus

it's raining, it's pouring,
it must be global warming

[Pointman]

In answer, some people delight in winning what is no more than verbal skirmishes while the war slips away from them unnoticed. Pygmies behind giant keyboards. They want their 15 minutes, pyrrhic or not.

Pointman

[Harry Passfield]

Bish: When Rice says, about Richard Tol:

I for example, particularly dislike that you regularly say things that you must know are not true

- he's gone beyond the pale. He has accused an esteemed scientist of knowingly lying. That is a new low.

[Bart]

Statistics always depend on a model. If the model is bad, then the statistics are bad.

Statistics are useful basically as a means of data compression - rather than looking at a whole data series, you can sum it up as behaving like this or that model with these parameters. They are not generally useful for discerning things you cannot see with your own eyes upon examining the data.

[David Jay]

"You have to read it now, don't you?"

Yup, you got me Bish...

[Richard Tol]

@Harry
It's low, but not new. Ken Rice has accused me of academic malfeasance before, but he prefers to smear my name over reporting me to the appropriate authorities.

[...and Then There's Physics]

Richard,

Ken Rice has accused me of academic malfeasance before, but he prefers to smear my name over reporting me to the appropriate authorities.

Once again untrue. I don't believe that saying things that are untrue on social media qualifies as academic malfeasance. Are you trying to prove me right, or do you not realise that you're doing it?

[Richard Tol]

@Ken
You may recall that you have also said that my comment on Cook, still forthcoming in ERL, contains falsehoods known to its author.

By the way, I write this in my professional capacity. In my personal capacity, I write about kids and music and football and stuff.

[...and Then There's Physics]

Richard,

You may recall that you have also said that my comment on Cook, still forthcoming in ERL, contains falsehoods known to its author.

I think I pointed out that if you published it with the claims as they were, it would be untrue things that you would know to be untrue. In fact, I may actually have said that you would be publishing claims that you couldn't know to be true. Either way, however, that still doesn't change that I have never accussed you of academic malfeasance. You do get this, right?

By the way, I write this in my professional capacity. In my personal capacity, I write about kids and music and football and stuff.

I'm really unclear as to what your point is. I think the obligation is on you to be careful as to what you say, not on me to care about the consequences of me pointing out that something you've said is untrue. Ultimately, you're responsible for your behaviour on social media, not me. That you happen to also be someone who works in this field is, as far as I'm concerned, utterly irrelevant. In fact, I find it a little bizarre that you would even suggest that it does.

Ultimately, you've said two things on this thread alone that are untrue, so if you really don't like me pointing this out, why not stop doing so. It can't be that hard. It's not as if this suggestion is unreasonable.

[Richard Tol]

@Ken
Publishing something that you know to be untrue is malfeasance.

[Bishop Hill]

Can you guys take this to the discussion forum please. The subject here is Radford Neal's post.

[...and Then There's Physics]

Richard,

Publishing something that you know to be untrue is malfeasance.

It could certainly be regarded as such, but if you're concerned about this, the onus is on you to not publish it, not on me to not point it out. However, let's clarify what I actually said:

if you publish this with the following statement

Cook et al. (2013) state that 12,465 abstracts were downloaded from the Web of Science, yet their supporting data show that there were 12,876 abstracts.

then you are publishing a statement that you do not know to be true. I would regard that as having publishing something dishonest. All you know is that there are missing identifying numbers, not missing abstracts.

So, I didn't claim that you would be publishing something that you would know to be untrue, I said you would be publishing a statement that you do not know to be true. There is a difference.

[...and Then There's Physics]

Sorry, the blockquotes got messed up. The second-to-last paragraph should have been part of the blockquote.