An Open Challenge to Dr. William Dembski

Posted 28 July 2009 by Ian Musgrave

Dear Dr. Dembski, in your recent post on Uncommon Descent "H.L. Mecken on the urge to save humanity", you quote approvingly from an article at the Investors Business Daily.

A new scientific paper [McLean, de Freitas, and Carter, 2009] says that man has had little or nothing to do with global temperature variations......Their research, published in the Journal of Geophysical Research, indicates that nature, not man, has been the dominant force in climate change in the late 20th century.

As a mathematician Dr. Dembski, you will certainly be able to effortlessly analyse the McLean, de Freitas, and Carter, 2009 paper (the full paper can be found here). Your challenge, Dr. Dembski, is to show exactly how this paper supports the statements in the Investor Business Daily article. Please pay particular attention to the smoothing and bandpass filtering analysis in your explanation. Oh, and you may also like to explain why the lead author John McLean states:

The paper by McLean et al (JGR, 2009) does not analyse trends in mean global temperature (MGT); rather, it examines the extent to which ENSO accounts for variation in MGT.

and what his statement means for the Investors Business Daily piece. You may also wish to consult these articles at Open Mind, Only in it for the Gold and More Grumbine Science, which do an in-depth analysis of the paper. Given your interest in the role of science in policy decisions, this challenge is a valuable opportunity for you to use your expertise to show how this paper refutes many decades of climate research.

50 Comments

harold · 28 July 2009

Investor's Business Daily just lost any chance of ever selling an issue or other product to this investor.

It's not just the fact that a scientific article is misrepresented, and then that single misrepresented article incorrectly suggested to be a contradiction of all the other literature to date on human contribution to climate change.

It's the childishly imbecilic tone. "Maybe the only place it's getting hotter is in Al Gore's head".

I'm not likely to want investment-related information if it comes from a source that lets that kind of idiocy run rampant. (For the record, I have also hesitated to read the WSJ of late, even though it has a tradition of good reporting away from the editorial page, to a much greater extent that IBD.)

Those who applaud themselves for their "hard headed" interest in business, yet who also follow the right wing cult right into outright science denial, are especially deluded. You're the exact opposite of hard headed - you're complete suckers.

I've noted over and over again that, logically, if there is even a moderate chance the human activity is contributing to severely unfavorable climate change, human activity needs to be modified.

KP · 28 July 2009

Couple of questions since I am not a climate scientist (IANACS?):

1) The "smoothing and bandpass filtering" thing escaped me. I know it has to do with the fact that the authors used the first derivative of both SOI and temperature deviation ("anomaly"). Can you or someone give a quick explanation?

2) In the RealClimate comments where McLean explicitly says we did not measure MGT, it seems like the moderator and other commentors are being unusually hard on the authors. My understanding is that the paper shows that ENSO is strongly linked with variance around the mean, but doesn't necessarily account for trends in the mean. McLean says nothing more, nothing less in the comment string. Am I missing something? It just seems like the model incorporates ENSO as a driver of "anomaly" better than previous models. Otherwise not an especially high-impact paper and definitely overblown by IBD. What a rag. To be used for lighting the barbecue or lining the feathered-dinosaur cage, at best.

Brian D · 28 July 2009

KP said: Couple of questions since I am not a climate scientist (IANACS?): 1) The "smoothing and bandpass filtering" thing escaped me. I know it has to do with the fact that the authors used the first derivative of both SOI and temperature deviation ("anomaly"). Can you or someone give a quick explanation?

Robert Grumbine explains this fairly well at More Grumbine Science; the link is in the article.

2) In the RealClimate comments where McLean explicitly says we did not measure MGT, it seems like the moderator and other commentors are being unusually hard on the authors. My understanding is that the paper shows that ENSO is strongly linked with variance around the mean, but doesn't necessarily account for trends in the mean. McLean says nothing more, nothing less in the comment string. Am I missing something? It just seems like the model incorporates ENSO as a driver of "anomaly" better than previous models. Otherwise not an especially high-impact paper and definitely overblown by IBD. What a rag. To be used for lighting the barbecue or lining the feathered-dinosaur cage, at best.

It's sort of a combination of what the paper says and does not say, and what the press release and resulting responses claim it says. Take a look at the Only In It For The Gold link for more information on this.

KP · 28 July 2009

Brian D said: It's sort of a combination of what the paper says and does not say, and what the press release and resulting responses claim it says. Take a look at the Only In It For The Gold link for more information on this.

Yeah, I haven't gone to those other links yet, but I will. Also, I just noticed this open can of worms at the end of the McLean et al. abstract:

That mean global tropospheric temperature has for the last 50 years fallen and risen in close accord with the SOI of 5-7 months earlier shows the potential of natural forcing mechanisms to account for most of the temperature variation.

Robert Grumbine · 28 July 2009

KP said: Couple of questions since I am not a climate scientist (IANACS?): 1) The "smoothing and bandpass filtering" thing escaped me. I know it has to do with the fact that the authors used the first derivative of both SOI and temperature deviation ("anomaly"). Can you or someone give a quick explanation? 2) In the RealClimate comments where McLean explicitly says we did not measure MGT, it seems like the moderator and other commentors are being unusually hard on the authors. My understanding is that the paper shows that ENSO is strongly linked with variance around the mean, but doesn't necessarily account for trends in the mean. McLean says nothing more, nothing less in the comment string. Am I missing something? It just seems like the model incorporates ENSO as a driver of "anomaly" better than previous models. Otherwise not an especially high-impact paper and definitely overblown by IBD. What a rag. To be used for lighting the barbecue or lining the feathered-dinosaur cage, at best.

I go on at more length at my blog, enough so that readers without any math background to speak of can (I hope) understand the time series issues involved. In briefer form here, the authors apply 2 filters to their data. First, they make a 12 month running average (and never say why, nor defend why it's 12 rather than 24 or 30 or ...). That, to be musical, suppresses the treble. Then they take the differences 12 of the data 12 months apart. That suppresses the bass. Having suppressed both the treble and bass, they've built a 'bandpass filter' -- same as you can on a stereo equalizer by shutting down the high and low frequencies, and jacking up the mid-range. That's kind of your question 1. Question 2 turns on a different matter. Namely, their conclusion makes claims about _all_ variability. Trends, long period variations, short period variations, are all part of the variability. They have not studied all variability, only variability that passes through their filters. So their claim here is simply false. They've explicitly excluded much of the variability, so cannot conclude about all variability. (It's explicit if you know as much about time series as is in my post that's linked to, and the introduction to time series that I wrote last Friday.)

Pete Dunkelberg · 28 July 2009

Stephen Meyer performs tonight!

He will be on Coast to Coast from 11 pm until 2 am! Meyer will "discuss recent discoveries in cell biology which support intelligent design and reveal that digital computers and living cells are operating on the same principles." He's uncommonly brilliant on this sort of thing.

If the program's web page doesn't convince you of something, I just don't know what to say.

Mike Elzinga · 28 July 2009

KP said: Couple of questions since I am not a climate scientist (IANACS?): 1) The "smoothing and bandpass filtering" thing escaped me. I know it has to do with the fact that the authors used the first derivative of both SOI and temperature deviation ("anomaly"). Can you or someone give a quick explanation?

The Open Mind critique addresses what comparing derivatives (rates of change) do to the data. If the temperature over the relatively short period (in geologic time) can be represented by a function of time plus a linear variation (i.e., a first order Taylor series on top of the all the other fluctuations) we have something like T(t) = f(t) + at, where f(t) contains all the fluctuations and the a is just the coefficient of the linear approximation to the overall global increase in temperature. Taking the derivative with respect to time simply leaves dT/dt = df(t)/dt + a. So all you are looking at are the rates of change of temperature offset by a constant. If the constant is small, which it is, plotting the derivatives of the temperature change will bury the constant in the data.

My understanding is that the paper shows that ENSO is strongly linked with variance around the mean, but doesn’t necessarily account for trends in the mean. McLean says nothing more, nothing less in the comment string. Am I missing something? It just seems like the model incorporates ENSO as a driver of “anomaly” better than previous models.

The point of filtering out other variations in order to concentrate on a particular variation is to get some idea of how much it contributes to the total variation. In the language of Fourier analysis, taking the Fourier transform of the data converts a time series to a frequency spectrum (after taking the absolute values squared of the transform). What his shows is the various periodic variations splayed out along a frequency axis with amplitudes along the vertical axis showing the relative contribution of each phenomena at their particular frequencies. All that the filtering does is chop off those parts of the frequency spectrum that are not of interest at the moment and concentrates the attention on the phenomena taking place at a particular frequency (in this case the ENSO). But when you attempt to do this type of analysis on the derivative of the data, the Fourier transform of the derivative a function goes like i x f times the derivative of the Fourier transform of the function itself, where f is the frequency (I’m leaving off multiples of 2 pi). The Fourier transform of a constant is an impulse function at zero frequency on the frequency spectrum. So, basically, the FT{dT/dt} becomes FT{df/dt + a} = (i x f)FT{f} + impulse(0). If you now filter by chopping off frequencies not of interest (that includes zero frequency) you throw away the data related to global warming. In short, the analysis is guaranteed to throw out the data on global warming; just the kind of analysis deniers love.

Mike Elzinga · 28 July 2009

My apologies for using f as a function and also as a frequency. Replace f(t) with g(t) and then df/dt becomes dg/dt.

fnxtr · 28 July 2009

Hmmm... another case of "we don't need to connect no steenking dots"? We'll see.

Ian Musgrave · 28 July 2009

G'Day All
Don't give too much of the analysis away. Dr. Dembski has to have something to analyse.

Mike Elzinga · 28 July 2009

Ian Musgrave said: G'Day All Don't give too much of the analysis away. Dr. Dembski has to have something to analyse.

Oops; sorry Ian. Need to resist the nerd thing. Mouth zipped.

DavidK · 28 July 2009

So has Dembski publicly, or privately, acknowledged and accepted the challenge, or is he remaining aloof like a good creationist?

deadman_932 · 28 July 2009

Erm...I tried a few diff. ways of trying to find any such .pdf at the site, and no go. It seems to have disappeared. Was it SATAN?!?!?!

Ian Musgrave · 29 July 2009

The PDf appears to have been taken down. Too many hits perhaps?

Mike Elzinga · 29 July 2009

Ian Musgrave said: The PDf appears to have been taken down. Too many hits perhaps?

I have a copy of it on my desktop so I could read it at my leasure.

Marion Delgado · 29 July 2009

Dr. Dembski is an expert - world class - on whether something is designed or random nature. Whether it's a bacterial flagellum or so-called global warming, he can immediately spot it. When he says that species are artifacts and climate change is random and normal, depend on it!

For the layman, climate change has no CSI. Hence, it's natural.

Steve P. · 29 July 2009

Poor Ian Musgrave.

You're asking Dembski to prove it is Nature (weather patterns) and not Nature (Man) that causes global warming.

Now showing live, the fight of the century: nature VS Nature.

Unless of course Man has a soul. Then we can have a rational debate.

Jim Ramsey · 29 July 2009

I can't help it.

and no fart noises, please

Kevin B · 29 July 2009

Jim Ramsey said: I can't help it. and no fart noises, please

Sorry. Are you suggesting that a) global warming causes an increase in wind strength?
b) global warming is caused by flatulent cows?
or
c) global warming is caused by US Federal judges who hand down judgements that annoy Dr Dembski?

Jim Ramsey · 29 July 2009

Kevin,

Yes, probably d) -- all of the above.

More substantially, I've noticed (well everyone has probably noticed) that Dr. Dembski tends to blow off serious questions that require actual work by playing the clown to his base.

It would be nice if he rolled up his sleeves and got serious.

carlsonjok · 29 July 2009

It would be nice if he rolled up his sleeves and got serious.

Ask and thou shalt receive!!

stevaroni · 29 July 2009

carlsonjok writes... Ask and thou shalt receive!!

For the love of God people - stop asking! If we receive much more of this, my blabbligcrap-o-meter will seize up and explode. Again. And I've just got the shop back in one piece after last weeks' Ann Gauger logic debacle.

Mike Elzinga · 29 July 2009

carlsonjok said:
It would be nice if he rolled up his sleeves and got serious.
Ask and thou shalt receive!!

Sheesh; “The Principle of Methodological Counterintuitiveness”, what a crock! If you can’t understand and do the science, just make up crap as fast as the rubes can eat it.

eric · 29 July 2009

carlsonjok said: Ask and thou shalt receive!!

Wow. It is just amazing how flagrantly Dembski quote mines. One Wikipedia sentence states that global cooling is scientifically bogus and that the public got confused because there was a slight cooling in temperature for a few years. Dembski takes that sentence as evidence for global cooling. Truly the man has no shame.

fnxtr · 29 July 2009

Mike Elzinga said:
carlsonjok said:
It would be nice if he rolled up his sleeves and got serious.
Ask and thou shalt receive!!
Sheesh; “The Principle of Methodological Counterintuitiveness”, what a crock! If you can’t understand and do the science, just make up crap as fast as the rubes can eat it.

Yeah, that hurt to look at, like a road accident. So if science and scientists are all evil and wrong and stuff, is Wee Willie going to turn in his Dr. Dr. papers? I liked the Libby Purves story linked in the sidebar, but wonder why they put it there. Sure it slags Dawkins, but it also says Christians can accept evolution. Maybe they never read it, they just saw it said something bad about Dawkins. Oops, I guess I just let the cat out of the bag, didn't I.

Ian Musgrave · 29 July 2009

carlsonjok said:
It would be nice if he rolled up his sleeves and got serious.
Ask and thou shalt receive!!

Hmm, in reply to commentators who say you have misrepresented climate science, quote an article which proves you did misrepresent climate science and then make stuff up. Well, what did we expect (and he's probably doing it just to stir scientists up anyway). What is more interesting The rest of the article is just more science envy. And the findings of science are counter intuitive. Well, yes, most of them are. Yes, some scientists get big money to make things like the Hubble telescope or the Large Hadron Collider (when they fix the soldering, sheesh) to discover things that warp our brain. Let's the DI propose their equivalent of Hubble, The Human Genome Project or the LHC with the data to back their proposal up.

Wheels · 29 July 2009

Hell, a bunch of things that are likely as useful as the Hubble or the Black Hole Factory (just you wait!!!11!) don't get any money anyway.

Kevin B · 30 July 2009

Ian Musgrave said: Let's the DI propose their equivalent of Hubble, The Human Genome Project or the LHC with the data to back their proposal up.

Most of the DI's constituency probably couldn't tell the difference between a Genome and a tacky garden ornament and think that the LHC won't work because a hadron is a duck-billed dinosaur.

eric · 30 July 2009

Ian Musgrave said: Let's the DI propose their equivalent of Hubble, The Human Genome Project or the LHC with the data to back their proposal up.

Any proposal review committee should also ask the DI what they have accomplished with the roughly $1million/year (for many years) they claim on their tax forms they've been spending on research. Even the a good proposal should be looked at with suspicion if the proposer has a track record of constant failure.

harold · 30 July 2009

Mike Elzinga -

Need to resist the nerd thing.

Actually, the nerd thing is greatly appreciated by some of us. Those of us with a biomedical education were all required to at least pass basic physics, general chemistry, calculus, and statistics to get our degree. Some of us held our noses and slogged through it, I'm sure. In many cases we had the option of "for the life sciences" courses. On the other hand, many of us enjoyed the stuff and take an interest in the fundamentals of other branches of science.

Mike Elzinga · 30 July 2009

harold said: Mike Elzinga -
Need to resist the nerd thing.
Actually, the nerd thing is greatly appreciated by some of us.

Thanks, Harold. Ian’s little reminder about wanting to let Dembski stew on it gave me a little tweak that perhaps I may not be aware when I have gone overboard in explaining something. On the other hand, I certainly appreciate the explanations of experts in other areas, especially the biology related areas.

Argon · 31 July 2009

Does Dembski apply the his 'Explanatory Filter' to global warming? Is it intelligently designed?

Mike Elzinga · 31 July 2009

Argon said: Does Dembski apply the his 'Explanatory Filter' to global warming? Is it intelligently designed?

Dembski’s filter appears to be very simple; if it isn’t ID, it’s wrong. (Or, in Dembski language: if it is the set theoretic complement of ID, then it is the set theoretic complement of correct.)

phantomreader42 · 31 July 2009

Argon said: Does Dembski apply the his 'Explanatory Filter' to global warming? Is it intelligently designed?

I don't think Dembski has ever consistently applied his "explanatory filter" to anything. It's just an obfuscation trick.

Marion Delgado · 2 August 2009

As you know I work in the Real Science division of the Discovery Institute. Last year, I had a shock when I applied the explanatory filter to Dr. Dembski's work, and it came out as a product of random mutation and natural selection, not intelligent typing.

Then i realized it had been more than 8 weeks since my last beating and confession, so I burned my results and turned myself in to Philip Johnson personally. I only want for the rest of you to feel the relief I feel right now. Problem solved, thanks to methodological supernaturalism.

Rilke's Granddaughter · 2 August 2009

Dembski has never actually applied his "explanatory filter" to anything. Ever. It's one of the funniest things about him - having invented this philosophical absurdity, he's never bothered to show that it can even be used.

If Dembski weren't such a bitter, nasty man - he'd be quite funny.

eric · 3 August 2009

Rilke's Granddaughter said: Dembski has never actually applied his "explanatory filter" to anything. Ever.

That may or may not be true. Behe and he claim that they've applied all of their baloney to three systems (mammalian immune, blood clotting, flagella) and shown those three to be complex specified information. However they never show their work. This is actually worse than never trying to apply it at all. Someone who thinks they've got a new technique for curing cancer and doesn't do anything about it is just crazy. But someone who claims they already have the cure in hand and refuses to share it is crazy AND malicious.

fnxtr · 3 August 2009

So, Dembski has the data to prove his point, but refuses to share it?
That's a good 20 points right there.

http://math.ucr.edu/home/baez/crackpot.html

Toidel Mahoney · 3 August 2009

Rilke's Granddaughter said: Dembski has never actually applied his "explanatory filter" to anything. Ever. It's one of the funniest things about him - having invented this philosophical absurdity, he's never bothered to show that it can even be used. If Dembski weren't such a bitter, nasty man - he'd be quite funny.

I have applied the explanatory filter to a cockroach, a cat, and my brother. My brother had the highest level of CSI, the cat was second, and the cockroach was last. I have calculated the exact amount of CSI difference necessary to define a Baramin. However, I would like to try it on a real evolutionist like PZ Myers in order to understnd if chronic sodomy can reduce the level of CSI.

phantomreader42 · 3 August 2009

Toidel Mahoney said:
Rilke's Granddaughter said: Dembski has never actually applied his "explanatory filter" to anything. Ever. It's one of the funniest things about him - having invented this philosophical absurdity, he's never bothered to show that it can even be used. If Dembski weren't such a bitter, nasty man - he'd be quite funny.
I have applied the explanatory filter to a cockroach, a cat, and my brother. My brother had the highest level of CSI, the cat was second, and the cockroach was last. I have calculated the exact amount of CSI difference necessary to define a Baramin. However, I would like to try it on a real evolutionist like PZ Myers in order to understnd if chronic sodomy can reduce the level of CSI.

And since you won't show your work, it's obvious you're just another lying creationist moron. Put up or shut up. Show your work, offer actual evidence, or go fuck yourself.

DS · 3 August 2009

TM wrote:

"I have applied the explanatory filter to a cockroach, a cat, and my brother. My brother had the highest level of CSI, the cat was second, and the cockroach was last. I have calculated the exact amount of CSI difference necessary to define a Baramin."

Well I have applied the unexplanatory filter to various things as well and the results have certainly been instructive. The CSI index for creationists turns out to be only 85 but for evolutionary biologists it is 163. For a cockroch it is 105 and for a rock it is 95. Seems as if it give the same results as an intelligence test. Coincidence? I don't think so.

fnxtr · 3 August 2009

I have calculated the exact amount of CSI difference necessary to define a Baramin.

Consider your chains yanked, folks.

Stanton · 3 August 2009

fnxtr said:
I have calculated the exact amount of CSI difference necessary to define a Baramin.
Consider your chains yanked, folks.

So what is the appropriate response to such maliciously asinine non sequitors? Light applause and a cookie?

Henry J · 3 August 2009

That would probably have as much, and as little, chance of getting through as anything else.

Henry J

Dan · 3 August 2009

Stanton said:
fnxtr said:
I have calculated the exact amount of CSI difference necessary to define a Baramin.
Consider your chains yanked, folks.
So what is the appropriate response to such maliciously asinine non sequitors? Light applause and a cookie?

Laughter and merriment.

Toidel Mahoney · 4 August 2009

DS said: TM wrote: "I have applied the explanatory filter to a cockroach, a cat, and my brother. My brother had the highest level of CSI, the cat was second, and the cockroach was last. I have calculated the exact amount of CSI difference necessary to define a Baramin." Well I have applied the unexplanatory filter to various things as well and the results have certainly been instructive. The CSI index for creationists turns out to be only 85 but for evolutionary biologists it is 163. For a cockroch it is 105 and for a rock it is 95. Seems as if it give the same results as an intelligence test. Coincidence? I don't think so.

The CSI index can only in principle be defined two ways. The traditional way is between zero and one, and the other way would be between one and infinity. (One is maximum CSI in either case.) It appears you have chosen the latter. Because of this, the score of 85 is the highest level listed, and the score of 163 is the lowest.

Kevin B · 4 August 2009

Toidel Mahoney said:
DS said: TM wrote: "I have applied the explanatory filter to a cockroach, a cat, and my brother. My brother had the highest level of CSI, the cat was second, and the cockroach was last. I have calculated the exact amount of CSI difference necessary to define a Baramin." Well I have applied the unexplanatory filter to various things as well and the results have certainly been instructive. The CSI index for creationists turns out to be only 85 but for evolutionary biologists it is 163. For a cockroch it is 105 and for a rock it is 95. Seems as if it give the same results as an intelligence test. Coincidence? I don't think so.
The CSI index can only in principle be defined two ways. The traditional way is between zero and one, and the other way would be between one and infinity. (One is maximum CSI in either case.) It appears you have chosen the latter. Because of this, the score of 85 is the highest level listed, and the score of 163 is the lowest.

But DS explicitly said that he was using the unexplanatory filter. This differs from Dembski's EF in that there is no pretence that the numbers obtained are anything except arbitrary values plucked out the air to support the point of view being argued. In any case, I think that the correct way to express EF values is as 10 times the log10 of the value, so that they can be quoted as deciBils.

Stanton · 4 August 2009

Toidel Mahoney said: The CSI index can only in principle be defined two ways. The traditional way is between zero and one, and the other way would be between one and infinity. (One is maximum CSI in either case.) It appears you have chosen the latter. Because of this, the score of 85 is the highest level listed, and the score of 163 is the lowest.

Please demonstrate how you used the Explanatory Filter to calculate CSI numbers, and please show your math, too. Otherwise, you're Lying for Jesus, which, last I heard, is a mortal sin that Jesus particularly loathes.

DS · 4 August 2009

TM,

Right. And just as soon as you show us your calculations, I'm sure we will all agree with your results.

As for me, I'll stick with the unexplanatory filter. At least it has units and you can measure the variance. You can also do statistics with it and determine confidence intervals. In every way it is superior to the so called explanatory filter.

Now why hasn't Dembski published any results in any scientific journals, or even shown a single calculation anywhere? Perhaps he just made the whole thing up to dupe the gullible rubes and couldn't fool any real scientists with his mamby pamby nonsense. Why on earth would the fig newton of delusional informational fantasies do such a thing? And why would he expect anything but ridicule from real scientists if he did?

Stanton · 4 August 2009

DS said: Now why hasn't Dembski published any results in any scientific journals, or even shown a single calculation anywhere? Perhaps he just made the whole thing up to dupe the gullible rubes and couldn't fool any real scientists with his mamby pamby nonsense. Why on earth would the fig newton of delusional informational fantasies do such a thing? And why would he expect anything but ridicule from real scientists if he did?

I'll take "Hard Truths of Liars for Jesus" for 1000, Alex!