Guest essay by Eric Worrall.
JoNova reports on a hilarious error in Lewandowsky’s paper “The Role of Conspiracist Ideation and Worldviews in Predicting Rejection of Science” . The calculated age of survey respondents has reportedly been skewed by one “outlier” who claimed to have been born in the palaeolithic, 32,757 years ago.
Raw data – line 607, http://websites.psychology.uwa.edu.au/labs/cogscience/documents/PLOSONE2013Data.csv
Lewandowsky was informed of this error over a year ago – and has reportedly done nothing to address this glaring problem with his calculations and conclusion.
According to Jo Nova;
“Lewandowsky, Gignac and Oberauer put out a paper in 2013 which was used to generate headlines like ‘Climate sceptics more likely to be conspiracy theorists’. The data sample is not large, but despite that, it includes the potential Neanderthal, as well as a precocious five year old and some underage teenagers too. The error was reported on Lewandowsky’s blog over a year ago by Brandon Shollenberger, then again by Jose Duarte in August 2014. Nothing has been corrected. The ages are not just typos, they were used in the calculations, correlations and conclusions. The median age was 43 but the mean age was a flaming neon 76. One wildly old person in the data skewed the correlation for age with nearly everything:
That one data point – the paleo-participant – is almost single-handedly responsible for knocking out all the correlations between age and so many other variables. If you just remove the paleo-participant, leaving the minors in the data, age lights up as a correlate across the board. Further removing the kids will strengthen the correlations.”
Worst of all, the bad data has apparently led to a demonstrably erroneous conclusion. According to Jose Duarte,
“This would be a serious problem in any context. We cannot have minors or paleo-participants in our data, in the data we use for analyses, claims, and journal articles. It’s even more serious given that the authors analyzed the age variable, and reported its effects. They state in their paper:
— “Age turned out not to correlate with any of the indicator variables.”
This is grossly false. It can only be made true if we include the fake data. If we remove the fake data, especially the 32,757-year-old, age correlates with most of their variables. It correlates with six of their nine conspiracy items, and with their “conspiracist ideation” combined index. It also correlates with views of vaccines – a major variable in their study. See the graph below.”
http://www.joseduarte.com/blog/how-one-paleo-participant-can-change-the-outcome-of-a-study
Lets hope that Lewandowsky finally takes action to correct the error, and amends the erroneous conclusion of his paper, which is inferred from Lewandowsky’s analysis of the grossly defective data. Of course, while he’s at it, I’m sure we could suggest a few other defects with Lewandowsky’s work which he could correct.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
It is well known that ages ending with 0, and 5 are most common in census data of rural communities (see http://ijphjournal.it/article/download/5630/5372)
The age data in this paper had a different problem, they were all synthetic. The digits 0,1,2 & 4 were used 25% times more than the others. In a normally distributed population of 1000 samples you should not expect to find differences more than 1-2%. This systematic overuse of these digits is a consequence of how digits are laid out on a typical keyboard.
I am not suggesting that the authors had their fingers on the keyboard, but who knows?
Out of a total of 1000 “people” in the list.
So, is the 5 year old also a valid data point?
Do four 14 or 15 year old children know enough to provide “policy information” that is cited when trillions of dollars of government spending is being invoked?
I am not disputing what you are saying. But, just highlighting that the age data were not real but synthetic.
You cannot have sampled a bunch of people in 30, 31, 32, 34 years of age but miss the 33 years old ones in a truly random pickings.
However, if your fingers on the left side of the numeric keypad, that is what you’ll end up with.
Lew just made the age data up, god knows he used the same trick in the other columns too.
8.5% of the data are ages with the same two digits. Not too far off of 10%.
The subjects were provided by the outfit who hosted the survey. According to the paper, they were chosen (obviously by a computer program) to satisfy certain requirements:
I doubt that such selections would be made by random methods which could account for the digit oddities in the sample.
Thank you for this possible explanation. Yet, I doubt the randomization functions in these algorithms are so poor to provide a footprint of their algorithms.
Regardless, the data in question are the ages of the participants. I would think the order number, ID number or other function would be the output of their algorithm but not the age since the former would contained all the access information to the participants.
I should stress the randomness test of digits is a well validated tool to detect synthetic data. In this aspect, Lew has failed, the age data is synthetic.
Age would be one of the factors actually used in the selection process in order to create a “representative” sample of individuals so the ages would definitely not be a simple random sample selected from the population available.
The effect on the distribution of the digits would depend on the specific algorithm used.
Roman, you are misunderstanding the point.
The respondents should be completely random. You don’t filter people out based on age except to remove impossibilities. That’s throwing away good data for no reason. A representative sample is accomplished by creating age divisions and analyzing the groups, not by deleting data. Either they filtered people out based on their age or the data is fake. Either way,
I have been extremely skeptical about this study since the beginning since the anonymous design is optimum for trolling. Lew might not have made this data up, but I’m fairly sure that many of his respondents did.
This would also explain the Neanderthal. Someone either put in a stupidly high age or a “-10”. I’m thinking the latter as the relation to the top of the short-int value is unlikely to be random, but it’s so obscure that it’s unlikely as a deliberate joke.
If we could ask anyone 32k years ago if they are global warming d3ni3ers, and they would have just wondered what warmth was.
‘amends the erroneous conclusion of his paper,’
if he did that for all of them , would there be anything left ?
Lew by name , Lew paper by nature the only real question is 2 or 3 ply
That he is a ‘star ‘ of the climate ‘science’ area , along with Mann, tell us much about the area itself.
Clearly, one-ply … he keeps getting his finger in sh8 data.
Let me guess. The 32,757 year old survey participant’s name is Treebeard and his home is in Yamal.
It is obviously falsified: There is no note or explanation about the “missing days” caused by the Julian date Gregorian date calender change, nor the “year 0” error induced by the assignment of the AD/BC years in the first place. Thus, a person born 32,757 years ago has NOT loved 32,757 years since his birth.
/sarchasm – That gaping whole between a liberal and real life.
Are you saying s/he is a hater (or maybe picture that Jimbo’s provided has biased your view of her ability to find love), or is it just that the finger on your right hand next to right ponky is a little overactive? 🙂
The fact that so many other scientists tolerate this nonsense without laughing it off the stage is the real tragedy. Science is beginning to deserve the bad rap it’s getting these days. Publication and peer review are turning into a joke.
Have turned. For more examples, see essays A High Stick Foul, By Land or By Sea, and Burning NonScience in Blowing Smoke: essays on energy and climate.
In climate science, data is only bad, if it doesn’t support your conclusion.
Hypothesis: People I Dislike, and Who Disagree with Me, Are Just Really Awful You Guys
It’s the science equivalent of an ad for a Klan rally.
A lot of politicians are corrupt, a lot of auto mechanics are corrupt, and a lot of scientists are corrupt. The difference is the cultural bias against the third fact, which needs to end before that corruption comes under even vague control.
Not that the first two are under control… tilting at windmills, perhaps.
The Lew problem is worse if you read Jose Duarte’s post and Brandon Shollenberger’s comments at Lucia’s.
The web survey service Lew used has a simple dashboard for each input. For age, it is NOT an input birthday as some posters have presumed above. It is a simple input age, where the dash allows to exclude ( e.g. 80) respondents. Obviously, no setting was used and somebody either spoofed or goofed. Utterly irresponsible. Jose has the dashboard screen shot, since he uses the same service for his Ph.D.
But even worse is Lew not QCing the SI data in his own Excel spreadsheet. Shades of Phil Jones incompetence. Anyone bringing me a mode/ mean variance like that upon which Lews conclusions rest would ‘have been made available to the competition’ the same day. And some were.
Sadly, these days you’d just enhance the grant-gobbling potential of the ‘competition’ in the process.
reading comments at PLoS paper itself is enlightening. AW should update that the comment on blogosphere coverage by WUWT and other skeptic sites.
With the Duarte comment now there at the paper, any credentialed scientist who now uses/references this LewPaper has to be considered”warned” that there are serious problems with the data and thus any conclusions.
PLoS has known of the serious data issues, and should be demanding the authors address them or face editorial retraction.
Important observation. lew has known for a year via Brandon. lew and PLOS1 have been on notice since (IIRC) last August via Duarte.. The reason this and like posts are important is that in the absence of correction/retraction, they comprise a clear irrefutable expose of ‘climate science corruption’ in the core scientific literature.
Yes Rud, even Emily Litella was able to clear her mistakes with “Oh. I’m sorry. Never mind.”
Its no surprise to find Lew has not even got the basic design of the study wrong , after all all that matters is that he gets the results he needs , how he gets them does not matter a dam . Bristol University must be hopping like hell that that he does not come unstuck on their watch , while if I was one of his students I demand to judge by his own rubbish standards so I could put any old rubbish in and still pass.
Lewandowsky has discovered the secret to immortality!
He probably has made himself immortal, but not, I suspect, in the way he intended.
“The Role of Conspiracist Ideation and Worldviews in Predicting Rejection of Science”
Please fix the link:
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0075637
NOT
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0075637%20
(Please delete this message.)
Why bother checking the data when you have the conclusion you want?
I am a conspiracy theorist and also don’t adhere to the science behind the global warming claim. It’s a better smear to be stuck with than being a Fox News follower!
[But being accused of being a Fox News follower is a compliment. .mod]
Huh?
So please tell us what your theory is about the conspiracy against Fox News.
It’s a better smear to be stuck with than being a Fox News follower!
Explain, please!
You are all wrong. That survey participant was an alien from a planet circling a brown dwarf in a 12 hour orbit.
Looks like someone entered their birthdate , not their age.
March 27, 1957
Looks like a researcher was too stupid and lazy to check the quality of the data for the faux study.
If that turns out to be right, ‘Down to Earth’, then it took only about 80 comments, and less than 4 hours to get the answer. That seems very efficient to me.
Here is Lewandowsky’s older research subject being interviewed:
Just more evidence that Lewandowsky is the shaman of the cargo-cult sect which worships the great and powerful God of Con$piracies.
John
32,767 is the upper bound of a 16 bit integer. Not 32,757, but the similarity is striking.
This is for visual basic. 2 raised to the 16th is 65536.
btw, 32767 is the upper bound of a signed 16 bit integer.
Robust statistics are a wonderful thing. Papers that use non-robust statistical methods (e.g., least squares) should be asked why by their reviewers.
Marginally competent scientists everywhere who are focused on creating exaggerated climate are very thankful to the PLOS Journal and Lewandowsky for lowering paper publishing standards.
John
I’m pretty sure the 32,767 year old is an agw skeptic…after all, he/she was born in the interglacial and has seen the holocene temperatures pretty much the same throughout these last 10,000 years before the relatively recent industrialization.
Born before the interglacial.
My bad…meant ice age…
Actually she was born in the relatively mild MIS 3 interstadial, lived through the Late Glacial Maximum, through the Late Glacial Interstadial, the Younger Dryas and the whole Holocene (including the Climatic Optimum), and the only thing that ass Lewandowsky can think of asking her is if she thinks princess Di was assasinated or not.
I have to Laugh at Lews conclusions, about right wing conspiracy ideation.
Surely an open minded researcher would have found that
Right wing views lead to life expectancy soaring
Male right wingers live 30 years longer than females
1 % of respondants were immune from ad-hom, crom-hom, add-on, nand add, ad-BOM insults
5 yr old Republicans are smarter than the average researcher
toorightmate
January 11, 2015 at 9:47 pm
I am nearing 70 YO and I am a pretty smart old codger.
The bloke that is 32,757 YO must be a bloody wizard.
Richard Tol (@RichardTol)
January 11, 2015 at 9:49 pm
This is deeply disappointing. Lew finds a veritable 32,757 years old, and all he can ask is “who killed JFK?”.
Baa Humbug
January 12, 2015 at 12:23 am
+1
Funnnnnnyyyyyy I can’t stop laughing. Thanx Richard.
(And here I thought economists didn’t have a sense of humor)
Yonniestone
January 12, 2015 at 4:45 am · Reply
John Maynard Keynes was one of my favorite comedians, until he became increasingly depressive…..
ref comments jonova
3,4,4,4,4,4,2,2,3,3,4,4,4,4,4,2,3,3,5,5,2,2,3,2,3,2,2,2,2,1,3,2,1,2,1,2,3,4,4,4,32757,2
Is this one of those “which number is not like the others” test?
How much time do I have to figure it out?
/grin