McIntyre on Lewandowsky’s Fake Correlation

Steve McIntyre takes Lewandowsky’s statistical screed to task and writes:

Lewandowsky’s most recent blog post really makes one wonder about the qualifications at the University of West Anglia Western Australia.

Lewandowsky commenced his post as follows:

The science of statistics is all about differentiating signal from noise. This exercise is far from trivial: Although there is enough computing power in today’s laptops to churn out very sophisticated analyses, it is easily overlooked that data analysis is also a cognitive activity.

Numerical skills alone are often insufficient to understand a data set—indeed, number-crunching ability that’s unaccompanied by informed judgment can often do more harm than good.

This fact frequently becomes apparent in the climate arena, where the ability to use pivot tables in Excel or to do a simple linear regressions is often over-interpreted as deep statistical competence.

I mostly agree with this part of Lewandowsky’s comment, though I would not characterize statistics as merely “differentiating signal from noise”. In respect to his comment about regarding the ability to do a linear regression as deep competence, I presume that he was thinking here of his cousin institute, the University of East Anglia (UEA), where, in a Climategate email, Phil Jones was baffled as to how to calculate a linear trend on his own – with or without Excel. At Phil Jones’ UEA, someone who could carry out a linear regression must have seemed like a deity. Perhaps the situation is similar at Lewandowsky’s UWA. However, this is obviously not the case at Climate Audit, where many readers are accomplished and professional statisticians.

Actually, I’d be inclined to take Lewandowsky’s comment even further – adding that the ability to insert data into canned factor analysis or SEM algorithms (without understanding the mathematics of the underlying programs) is often “over-interpreted as deep statistical competence” – here Lewandowsky should look in the mirror.

Lewandowsky continued:

Two related problems and misconceptions appear to be pervasive: first, blog analysts have failed to differentiate between signal and noise, and second, no one who has toyed with our data has thus far exhibited any knowledge of the crucial notion of a latent construct or latent variable.

In today’s post, I’m going to comment on Lewandowsky’s first claim, while disputing his second claim. (Principal components, a frequent topic at this blog, are a form of latent variable analysis. Factor analysis is somewhat different but related algorithm. Anyone familiar with principal components – as many CA readers are by now – can readily grasp the style of algorithm, though not necessarily sharing Lewandowsky’s apparent reification.)

In respect to “signal vs noise”, Lewandowsky continued:

We use the item in our title, viz. that NASA faked the moon landing, for illustration. Several commentators have argued that the title was misleading because if one only considers level X of climate “skepticism” and level Y of moon endorsement, then there were none or only very few data points in that cell in the Excel spreadsheet.

Perhaps.

But that is drilling into the noise and ignoring the signal.

The signal turns out to be there and it is quite unambiguous: computing a Pearson correlation across all data points between the moon-landing item and HIV denial reveals a correlation of -.25. Likewise, for lung cancer, the correlation is -.23. Both are highly significant at p < .0000…0001 (the exact value is 10 -16, which is another way of saying that the probability of those correlations arising by chance is infinitesimally small).

These paragraphs are about as wrongheaded as anything you’ll ever read.

Read the rest here at Lewandowsky’s Fake Correlation

About these ads

50 thoughts on “McIntyre on Lewandowsky’s Fake Correlation

  1. So Lewandowsky took garbage data, applied inept and inappropriate “statistical analysis” and ….. voila! He has a paper for Psychological Science to warm the cockles of every Alarmist heart.

    Meanwhile, down under in Lew-Lew Land, he has published NINE stream of consciousness blog articles blasting critics and failing to address the substantive issues.

    What a shame, “Professor” Lewandowsky, you are not contributing to understanding (well except by negative example of what not to do). Perhaps such self-immolation of a fervent CAGWarmist is educational somehow….

  2. I almost feel sorry for him…. A Psychologist trying to teach McIntyre about statistics….

    The Rule of Holes seems to apply here.

  3. In the same way that I would like to speak a foreign language,but know that I can’t get the hang of it, I find statistical analysis fascinating even though most of it goes straight over my head.

  4. This post highlights a major problem in climate science. We have a team of alarmists who are not expert at statistics but are using statistics to give governments around the world the ammo to demand crazy restrictions on mankind’s commercial activities (CO2 reduction); not to mention Billions upon Billions of tax dollars down the toilet.

    Why are they able to get these papers, based upon statistics, published without having a real statistician co-author? Could a Phil Jones type ever be trusted (no matter which side he is on) if we can’t trust his abilities in statistics?

  5. I understand that Climate Science proponents have a tendency to believe in left-wing conspiracy theories like 9/11 was faked by Republicans to jusitfy Gulf War II. Funny, no such question was included by loony Lew.

  6. The science of statistics? Lewandowsky’s “work” has nothing to do with science, but rather the old attempt of using statistics based on very small groups as an evidence for vastly larger groups, which pretty ridiculous and unscientific.

    Science and statistics are not synonymous.

    Discovery of a numerical discrepancy is not science. Accounting for that discrepancy in a reproducible manner is science.

  7. UWA used to be a very highly regarded institution once, but as with nearly all others, ideology has probably trumped education. I wasted a year there in the ’60s, but that was my fault …

  8. I think someone has fed Lewandowsky with some BIG statistics words which he thinks he can use to baffle the rest of the world with,
    As they say “A little knowledge is dangerous”.
    After Steve has finished with him I would like to see what Mr William Briggs has to say about Lewandowsky’s effort.

  9. The problem is not insufficient “statistical competence”. EVERY use of statistics is an attempt to compensate for lack of data, and expert use of stats is more dangerous than amateur.

    If you’re dealing with an actual natural phenomenon, you need to analyze it with actual measurements and (where possible) an actual experiment. Statistics are an admission of failure.

  10. Prof Lewandowsky complains that there has been little discussion of the concept of latent constructs. Plenty of description on the Net about what they are. In one invented example, using a collection of observations to suggest a factor-in-common, like using measured observations such as wealth, happiness in marriage, abundance of food, smiling often, giving gifts … all lumped together into a ‘quality of life’ which makes the stats analysis less cumbersome.

    Our first child started to bring home kindergarten finger paintings that were painted all in black. We researched the possible deep significance of this as a latent construct, found lots of psychological reasons to be worried. At the appointment with the Kindergarten Head, we were promptly told that the kindy had run out of all paint colours except black.

    The latent construct, used badly, can lead up a garden path.

  11. Psychologists are some of the most sophisticated statisticians in academia. Lewandowsky can do the sums correctly, but poses the wrong questions. He proves nothing, it is shallow. It is clear to me that he assumes all of his nominated questionaire “conspiracies” (and perhaps we need to examine what he means by that term) are false and irrational by definition and takes off from there. Wouldn’t it be prudent to provide some weighting to the relative craziness of each conspiracy in the questioniare when doing the stats? The fact is, every now and then hidden conspiracies are exposed; for example I believe the Iran-Contra affair is no longer a theory but clearly was a deliberately hidden conspiracy. Perhaps even the most fundamental question, why try to even look for an association between climate science scepticism and conspiracy theories. Does he assume that all sceptics think there is some sort of conspiracy going on? He’s sounding like the conspiracy theorist to me, all those bad sceptics are banding together in secret and doing bad things.

  12. “If your experiment needs statistics, you ought to have designed a better experiment”

    – Ernest Rutherford

  13. Hi Anthony. Have just read some puerile tweets from so called US “journalists” abusing you and your site. All I can say is take heart, you have major support out here for your work. You may like to know that our family (which includes a paleontologist, a marine biologist, a statistician and a chef) all look forward to reading the articles and discussing their various merits.You do a fantastic job to pull it all together so professionally.Cheers and best wishes from Oz. We think you are wonderful.

  14. This is the first time I have looked at a blog post by Lewandowsky and Oberauer, and so I don’t know if any one else has made the comment, But what on earth is Lewandowsky doing in creating a post on what he perceives to be a Sceptic V “Realist” attitude to interpreting time series data in a discussion which is supposed to be about detecting the signal from noisy data? His selection of what he perceives as sceptics traits in interpreting graphical data surely demonstrates his highly negative attitude to sceptics and therefore his professional unsuitability in producing and interpreting sceptics attitudes to climate science and moon landings. Coupled with his clear associations with Cooke and “Sceptical Science,” this should bring into grave doubt any conclusions that might have been gained from his survey results and in an honest world withdrawal should have been automatic.

  15. ” data analysis is also a cognitive activity”. & “informed judgment” are code words for “what ever I want the data to show, is what the data shows.”

    The sad state of climate scientists today [major facepalm].

  16. markstoval says:
    September 19, 2012 at 2:45 am
    This post highlights a major problem in climate science. We have a team of alarmists who are not expert at statistics but are using statistics to give governments around the world the ammo to demand crazy restrictions on mankind’s commercial activities (CO2 reduction); not to mention Billions upon Billions of tax dollars down the toilet.

    Why are they able to get these papers, based upon statistics, published without having a real statistician co-author?
    __________

    Because the “peers” who review the papers do not understand statistics either. Moreover, a “if it bleeeds (CAGW) it leads” mentality seems to be pervasive amongst Journal editors.

  17. This is a good example of politically motivated “subjective research (post modern science)” at it’s worst. I advise “true believers” to maintain some distance or go down the drain with him.

  18. Psychology is to science as Alchemy is to Chemistry.

    But it employs a lot of people who otherwise would be only qualified to ask folks if they would like fries with their order.

    At least Lew has a fall back career.

  19. I’m not a professional statistician, but I have used statistical techniques extensively in my research and know that it can be missused. Around 45 years ago I wrote an award winning paper on interpretation of experimental data that has been included and maintained in a handbook now in it’s third edition. http://books.google.com/books?id=8C7pXhnqje4C&pg=PA83&lpg=PA83&dq=fred+h.+haynie+statistics&source=bl&ots=tpEUeDItOc&sig=cfZfkW3QZd_3r-N4XyzleWkz_E4&hl=en&sa=X&ei=2clZULK5KIue8QS8mYGQAw&ved=0CDIQ6AEwBA#v=onepage&q=fred%20h.%20haynie%20statistics&f=false
    It presents the basics of using statistical techniques that Lewandowski does not seem to understand. Another way of finding that paper is to google Fred H. Haynie +statistics.

  20. With apologies to Humpty Dumpty: “The data mean precisely what I intend them to mean, neither more nor less”

  21. Chris B says:
    September 19, 2012 at 6:40 am
    Lews notes to self: 1.Work on understanding Stastics.
    2. Work on Acting.
    #, Use less starch.

    Yikes!

    What a prat! How can a scientist sceptical of AGW get his or her paper peer reviewed & published if other scientific prats elbow their papers out of the peer review system because they don’t want them aired in public. Therefore publishing said paper on the internet where it can be peer reviewed by millions & millions, sceptical & otherwise, expert or not, for free!! Seems far better to me than publishing in discriminating journals! Oh & what’s with the Lloyd Grossman impression?

  22. “The science of statistics is all about differentiating signal from noise.”

    The signal here is that we skeptics are noise and the CAGW crowd is the signal. Playing little insults to those who have the nerve to disagree with the settled science.

  23. http://www.shapingtomorrowsworld.org/about.htm.

    About Us

    Mission Statement

    From climate change to peak oil and food security, our societies are confronted with many serious challenges that, if left unresolved, will threaten the well-being of present and future generations, and the natural world. This website is dedicated to discussion of those challenges and potential solutions based on rigorous scientific evidence and objective scholarly analysis.

    Our goal is to provide a platform for re-examining some of the assumptions we make about our technological, social and economic systems. The posts on this site are generally written by domain experts, specialists and scholars with an interest in these problems and we hope they will generate informed and constructive debate. We will archive seminal papers and posts for future reference.

    Principals

    Prof Stephan Lewandowsky (School of Psychology, University of Western Australia)

    Prof Steven Smith (ARC Center of Excellence in Plant Energy Biology, University of Western Australia)

    Editorial Board

    Prof Glenn Albrecht (School of Sustainability, Murdoch University)

    Asst Prof Mark Edwards (Business School, University of Western Australia)

    Assoc Prof David Hodgkinson (Law School, University of Western Australia)

    Prof Carmen Lawrence (School of Psychology, University of Western Australia)

    Editorial Policy

    Shapingtomorrowsworld welcomes submissions from members of the public, for posting on our discussion board. However, all submissions will be reviewed by at least two members of the editorial board and acceptance cannot be guaranteed.

    Use the “contact us” button at the top to send us email or to submit a post.

    Comment Policy

    Our comment policy is rigorous. Although we welcome debate, we set very strict parameters on what we consider civil and substantive. If in doubt, a comment will be deleted. To facilitate the conditions for reasoned debate, comments can only be left upon registering with a valid email address.

    To avoid running afoul of our policy, we urge you to read the hints about commenting that you will receive upon registering for comments or read our policy now.

    The editors may approach commenters to turn their comments into blog posts if they find points in comments particularly pertinent.

    Funding

    Shapingtomorrowsworld was made possible by a grant from the Vice Chancellor of the University of Western Australia and by the support of the Institute of Sustainability and Technology Policy at Murdoch University.

  24. Besides the recognition of what is noise and what is signal, there is another serious problem in data analysis, the lack of sampling theorem of Nyquist-Shannon. When studying Paleoclimatology, more specifically ice cores, dating of the top and bottom of the samples have ranges 200-400 years (http://www.ncdc.noaa.gov/paleo/icecore/antarctica/domec/domec_epica_data.html). In the process of gas analysis, integrates the sample obtaining an average. This average not represents the extremes, ie if there has been an event Henrich the signal is filtered. If we compare the data from EPICA Dome C or Vostock with GRIP or NGRIP data it is evident (http://www.gfy.ku.dk/~www-glac/data/NGRIP_d18O_50yrs.txt).
    The problem is that VOSTOK or EPICA are taken as the record of the climate over the past 400 ky BP (# http://www.grida.no/publications/other/ipcc_tar/?src=/climate/ipcc_tar/wg1/072.htm fig223) and GRIP / NGRIP (samples ~ 20 years) only 100 ky BP.

  25. Colin Porter says:

    This is the first time I have looked at a blog post by Lewandowsky and Oberauer, and so I don’t know if any one else has made the comment, But what on earth is Lewandowsky doing in creating a post on what he perceives to be a Sceptic V “Realist” attitude to interpreting time series data in a discussion which is supposed to be about detecting the signal from noisy data?

    What he is doing is his schtick.

    In his Internet Survey paper, he takes fake data created by warmists pretending to be sceptics and uses it to draw unflattering conclusions about sceptics.

    The SKS blink chart is a fake chart created by a warmist pretending to be a sceptic and he uses it to make unflattering conclusions about sceptics.

    Its what he does.

    That, and call people “conspiracy theorist”.

    Typical witch doctor turned propagandist.

  26. “…Blog analysts have failed to differentiate between signal and noise…” –Stephan Lewandowsky

    Lewandowsky’s paper is 100% noise, 0% signal. It should never have been accepted for publication. But then, this IS climate science, the only hard science with a proud tradition of non-replicability.

  27. Chris B says: “http://www.shapingtomorrowsworld.org/about.htm.
    About Us; Mission Statement….”

    Should have been called “shagging tomorrows world.”

  28. First this is not science it is pseudoscience. Second mathematics is not science it is a tool that is used in science and lots of other applications too. None of this is science because it is founded on a priori reasoning. It is very had to believe an educated person could be so stupid unless that person had some conspiratorial agenda. It is sad and unfortunate that so much time and resources are being wasted by this foolishness. This kind of junk is often fostered by the extreme ideologists on any and all sides of a question as a smoke screen to obfuscate the truth of the matter.

  29. I agree this Lew’s paper is garbage, but even if it wasn’t.. who cares? What would it prove either way that some skeptics believe in pop culture conspiracies? Mann et al. apparently believe in the “big oil funding of skeptics” conspiracy. I have warmist friends that believe in the “grassy knoll” and even that the moon landing was staged. And what about all these people’s beliefs in one religion or another? Could any correlation between those observations prove or disprove the hypothesis of AGW or anything else for that matter?

    And remember, correlation is not causation.

  30. I remember back in 2005/6 a co-worker at Boston Scientific, who had worked for General Mills for 30 years doing D.O.E. (Design of Experiments) asked me why we used TABLES for computing the “F distribution” and then finding “p” values for significance.

    I brought in my most involved Statistics book which gives all the analytical formulas for the various distribution functions which are commonly used. (Look here for the F distribution…

    http://en.wikipedia.org/wiki/F-distribution)

    I laid out the equation, used MATHCAD to do both symbolic reduction and numerical analysis.

    I took some data he had, and put it into the correct forms and derived a “p” factor of .96…he said, “That’s wrong…it should be less than .05!” Then I pointed out the 1 MINUS in the formulation, and he realized we’d gotten almost EXACTLY the value we had from the tables.

    I then pointed out the time. An hour and 1/2 from when we had started. “Charlie, how long does it take you to us MiniTAB or the Tables to get to this value?” Charlie said, “About 2 minutes.”

    And this took us??? Charlie thought, “Wow, and hour and 1/2…!!!”

    I lost my best book for a MONTH as Charlie borrowed it to “bone up” (the book discusses the BASIS for the distributions, the derivation of the formulas, and the proper circumstances to use them…)

    When you start doing “closed form” integrals on Gamma functions, Error functions and the like, it gets pretty hairy!

    What tells me that LEW would be totally LOST within MINUTES if I sat down with him and discussed the TRUE BASIS for the statistics he THINKS he understands?

    Max

  31. Lewandowsky needs Auto Tune to rid himself of the pompous nerd affect.

    Is there a Doctor in the house who can decipher the micro-emotions Lew is machine-gunning at the camera in his bizarre clips?

  32. Any reasonable person, on hearing that someone with Steve McIntyre’s proven track record had shredded his statistics, would go very, very quiet, and start a frantic correspondence with Steve to get it right.

    That’s all.

  33. Dana & the CSRRT?

    Let’s take the CSRRT’s mission as rapid response to all things skeptic, but not credible response; which probably explains Dana’s incompetent rapid responses wrt PBS hosting Anthony.

    John

  34. Lewandowsky continued:

    “We use the item in our title, viz. that NASA faked the moon landing, for illustration.”

    No Lewandowsky, the conspiracy theorists are on your side. The warmie side. Those that are most likely to believe in Roswell UFO’s, grassy knoll shooters, faked moon landings, 9/11 WTC inside job by Cheney and the Jews, will also believe in Global Warming from the magic molecule. This is Roseanne Barr and occupy Wall Street IQ level. They are Socialist malcontents, hardwired in their DNA to accept government control. They are your target audience.

  35. A fact frequently becoming apparent in the climate arena, is that carrying the title ‘professor’ is often over-interpreted as implying deep scientific competence.

  36. An interesting thing about many CAGW zealots (as opposed to those who have one way or another been honestly convinced that CAGW is a present danger), is they are, almost to a person, so naive about how the internet works. They appear to have no idea that when they are trying to baffle, give a snowjob or present simplistic or clumsy theory to people or pontificate about issues, techniques, science, data, logic,…on a blog, they are presenting their stuff and reputations to, among the many, some highly intelligent, top- of- the- line practioners who dwarf the expositor’s comparatively puny capabilities. They clumsily hide declines, fudge data, hide data, dispose of data, manipulate data, cherry-pick and amaze themselves with their analytical prowess in front of giants. This correlation between zealotry/hubris and naivete is more telling of the person than their theory.

  37. I don’t know why we are all getting uptight. This seems like a very typical climate consensus paper as far as I can tell. Think about it.

Comments are closed.