More on the Mashey-Wegman issue. Part 1 is here
Guest post by Thomas Fuller
John Mashey has written one of those conspiracy theory plots full of colored dots and ink, accusing Edward Wegman of plagiarizing the work of Raymond Bradley and others in his key report to Congress showing the deep flaws in the paper Bradley wrote with Michael Mann and Malcolm Hughes regarding the Hockey Stick Chart.
Bishop Hill has so far said what needed to be said most succinctly:
there are two possibilities in play:
Wegman et al are guilty of plagiarism; short-centred principal components analysis is biased and can produce hockey sticks from red noise
Wegman et al are not guilty of plagiarism; short-centred principal components analysis is biased and can produce hockey sticks from red noise.
There is a bit of meat with John’s Mashed potatoes, sadly, and I’ll get to it in a minute. But let’s get rid of the twisted conspiracy theory garbage first.
In yesterday’s article I noted that it was extremely strange that Mashey would make as his first point the fact that the Congressional sub committee that commissioned Wegman turned over the results of the work they had done prior to Wegman starting his analysis. This is absolutely normal and uncontroversial, but Mashey writes as if it’s evidence of conspiracy, something he seems to find everywhere he looks.
It is also bizarre that Mashey thinks it wrong that works mentioned in the bibliography to Wegman’s report are not cited. This is clearly evidence that Mashey doesn’t understand very much at all about how anything really works. As anybody familiar with publishing knows, the reason a bibliography exists is to show the reader what the author read, precisely because the works may not be cited in the text. But again, this becomes black helicopter conspiracy for Mashey.
But there is a little meat with Mahsey’s potatoes. Please continue.
About half the plagiarization accusations in Mashey’s paper don’t even concern the Wegman report, targeting the recent McShane Wyner paper and dissertations and other collegiate work done by some of Wegman’s associates. Given that the title of Mashey’s report is Strange Scholarship in the Wegman Report, I don’t really see what criticism of other work is doing there. I guess it’s all there to prove a grand conspiracy.
But there is a little meat with Mahsey’s potatoes. Please continue.
In case you think I’m making this conspiracy stuff up, just read pages like 103 of Mashey’s report. It isn’t about Wegman at all. It’s about McShane Wyner,
In a report accusing Edward Wegman of plagiarism, Mashey writes this about an unrelated paper published 4 years later:
“MW, p.2, Paragraph 3
―On the other hand, the effort of world governments to pass legislation to cut carbon to pre-industrial levels cannot proceed without the consent of the governed and historical reconstructions from paleoclimatological models have indeed proven persuasive and effective at winning the hearts and minds of the populace. Consider Figure 1 which was featured prominently in the Intergovernmental Panel on Climate Change report (IPCC, 2001) in the summary for policy makers1. The sharp upward slope of the graph in the late 20th century is visually striking, easy to comprehend, and likely to alarm. The IPCC report goes even further:
‖<B> ―world governments … consent of the governed‖
When one Googles the 6 words above,50 many hits espouse strong conservative/Libertarian political views. Those are fine in the political arena, but not in statistics papers people expect to be credible. From past experience, 51 strong political/ideological beliefs can cause a few physics PhDs to ignore basic laws of physics.”
But there is meat to go with Mashey’s potatoes.
Wegman’s report examined the relationships between the very small community of scientists working in and around the field of paleoclimatology. This is because after Mann’s Hockey Stick chart began receiving criticism, a flurry of papers were suddenly published supporting his results–and they all received prominent attention and temporarily saved Mann’s reputation. But, as Steve McIntyre pointed out, these scientists were all closely connected to Michael Mann, being co-authors, co-bloggers, mentors and advisers of his. Worse, they used the same suspect data and the same discredited analysis techniques.
Wegman formalized an examination of Mann’s compatriots using a relatively new discipline called Social Network Analysis. And in his report to Congress, when Wegman explains what Social Network Analysis is, he copies someone else’s introduction of the science and doesn’t attribute it at all. (It looks like Wikipedia is copied, which means that someone else probably got copied to get it into Wikipedia. There is even the slight chance that Wikipedia copied Wegman, considering dates and such, but that would be just too delicious, so it probably didn’t happen that way.)
So it looks like Wegman and his team did something wrong. They used someone else’s description of social networking analysis and didn’t credit them.
But as near as I can tell, that’s it. And to put their error into perspective, let’s look at how other independent professionals describe social networking analysis without attributing each other:
Social network analysis is concerned with understanding the linkages among social entities and the implications of these linkages. The social entities are referred to as actors that are represented by the vertices of the graph.
Wegman Report
It is concerned with understanding the linkages among social entities and the implications of these linkages. The social entities are referred to as actors that are represented by the vertices of the graph. Most social network applications consider a collection of vertices that are all of the same
WK Sharabati
A social network analysis must also consider data on ties among units” … The entities in digraphs are called nodes and the relations are ….. The density measure describes general level of linkage among the actors in the community
Kilkenny and Nalbarte
A relation is represented as a linkage or a flow between these … relationships among social entities. In Social Network Analysis we can ….. network is referred to this focal person, and every relation is reported by the ego. …. “nodes” of the graph, and the “ties” between actors in the network become “lines”
F Martino
Social network analysis views social relationships in terms of network theory consisting of nodes and ties. Nodes are the individual actors within the networks, and ties are the relationships between the actors.
Wikipedia
Social network analysis [SNA] is the mapping and measuring of relationships and flows between people, groups, organizations, computers, URLs, and other connected information/knowledge entities. The nodes in the network are the people and groups while the links show relationships or flows between the nodes.
Valdis Krebs
Thomas Fuller http://www.redbubble.com/people/hfuller
According to an update of the USA Today article, GMU has begun a formal investigation. Pass the popcorn.
For all of you who say this is not serious, well GMU seems to think it is.
There were a lot of people involved in producing the Wegman et al. report:
@Nick Stokes
“The real issue is that the Wegman report was supposed to represent the considered opinion of an eminent statistician, and so to carry weight. If it turns out to include stuff pasted from Wiki, people feel cheated. And it really doesn’t help to say that the pasting was done by junior interns.”
The text has no bearing on the content or conclusions of the report. Whether the information concerning SNA was in Wegman’s own words, an intern’s own words, or attributed to a 3rd party would have absolutely no bearing on the content, conclusions, or “weight” of the report. Therefore, it’s insane to conclude that this material was plagerized for credibility when attributing it would be easy and not change a thing.
I conclude that it’s a documentation snafu and much-ado-about-nothing. It’s not good, it’s a distraction, but it’s immaterial.
Hmmm – another example of climatological copying…
Bradley, 1999 – as quoted by Mashey “Many natural systems are dependent on climate; where evidence of such systems in the past still exists, it may be possible to derive paleoclimatic information from them.”
The Chen, Xing, 2008 “Many natural systems are dependent on climate; where evidence of such systems in the past still exists, it may be possible to derive paleoclimatic information from them.”
(i.e. a direct copy)
http://books.google.com/books?id=SEO_RyNDJ0gC&lpg=PR1&pg=PA50#v=onepage&q&f=false
Seems to me that copying is common in climatology.
A report for the government does not need to meet normal standards of original research or academic writing. Given that Wegman et al were not experts in paleoclimatology and were probably working under severe time pressure, he and his associates logically should have pasted together blocks of text from authoritative sources. However, under no circumstances (even in this blog) should ANY educated person ever paste a block of text copied from someone else without properly annotating the source of the text with a description such as: “The following information about XYZ was taken from: ‘Pasted text’ “. Using an ordinary citation without quotes is simply inappropriate when any block of text is copied and pasted.
One might speculate that blocks of text were copied and pasted by one of Wegman’s students without full annotation and possibly without realizing how the text would eventually be used. That text probably was edited by other authors who did not understand the source of some blocks of text. I’m skeptical that any investigation will show that Professor Wegman himself personally copied and pasted blocks of text and then (as senior author) let them be used inappropriately in the final version of his report. I suspect that some of his co-authors are going to be answering awkward questions about why they didn’t properly document text that was copied and pasted, whether they knew how that text would be used, and whether they were given an opportunity to correct their mistake before the report was issued.
Of course, possible plagiarism don’t invalidate the scientific conclusions of the Wegman Report any more than the investigations by the Attorney General of Virginia invalidate the scientific conclusions of MBH98 and MBH99. It just damages personal reputations.
ZT,
Seems like Bradley has some other people who are really violating his copyright AND making money off it. The question is will he go after them?
Lazar says:
October 11, 2010 at 4:00 am (Edit)
Steven Mosher,
“In full possession of the definitions you give I would still not consider it to be plagiarism.”
… so, using a highly non standard definition of ‘plagiarism’, you would not consider it to be ‘plagiarism’… is what you’re saying… ok
#########
no silly. as somebody who actually had to make the decision whether to bring cases forward and ruin a kid, I am saying that if one of my students did this i would not consider it to be plagiarism.
If we use the bright line of the standard much of what bradley wrote in his textbook would be plagiarism. The bright line of the standard exists to give prior notice as to what will count as plagiarism. You still dont get what I’m saying. Thats ok.
Nick Stokes says:
October 11, 2010 at 3:33 am (Edit)
Steven Mosher says: October 11, 2010 at 1:01 am
“No attribution that I can see”
I can’t see whether there is or not. Maybe US users get to see a more complete version. All I see is a tiny window with a fragment of the para – no context at all. Various people, including Glenn, say there’s a reference on the next page – I don’t know how it is connected to the quote.
###################
There’s no attribution prior to the section. Also, no quote marks.
the point is that Bradley’s work is a text book. I will bet if I look through his textbook I will find parts of it ( like the appendix on carbon 14 dating) where bradley has not followed the rules exactly. Where he is citing background material and forget to quote or paraphrase correctly. And we will find textbooks after bradley that dont get it right either. the wegman committee did not get it right. using other peoples material is tough work, especially when its just background in an area where you are not FLUENT. Because the commitee was not fluent in paleo and in SNA, they had to use the words of others.
Glenn says:
October 11, 2010 at 1:37 am (Edit)
Steven Mosher says:
October 11, 2010 at 1:01 am
You didn’t cite the source showing Bradley said that. And at the end of the section of Cuff and Goudie’s book on the next page after the quote, page 132 shows this:
“Bradley, R.S. Quaternary Paleoclimatology: Methods of Paleoclimatic Reconstruction. Boston: Allen and Unwin, 1985.”
############
Glenn: There are two issues: Attribution and quotation.
Wegman attributes Bradely, But he does not use quotation properly.
The oxford companion also fails to use quotation properly.
The bradley case against the wegman committee is extremely weak and I’ll suggest again to people that many, ( including bradley ) have done what the wegman committee appeared to do. They cited bradley properly, but they failed to indicate in the text that they used BOTH bradleys ideas and his words.
The attribution shows us the SOURCE, quote marks show us the kind of usage. People who attribute correctly but fail to indicate the level of usage are not intellectual thieves ( the worst of plagiarism). technically, its plagiarism, copyright violation, but not a good case.
In the SNA case they fail to cite or indicate the KIND OF USAGE. Thats worse.
The reason the DONT CITE the source is cause they are probably embarassed they used wikipedia and didnt read primary material. ( sounds like a grad student )
Bad.
Anyway, if people want to go after the best argument they will focus on the SNA stuff and the real reason why they didnt cite the source
It would be interesting to use social network analysis, which btw, is concerned with understanding the linkages among social entities and the implications of these linkages. The social entities are referred to as…. ah you guys know what I mean. It would be interesting to see if among Bradley’s incestuous connections he himself copied someone else’s definition of dendrochronology. Anyone up to checking out Bradley’s work for plagiarism? Now that would be rich!
Bradley needs to explain why the copying of text is justified when he does it (in papers A and B) but unreasonable when Wegman does it, and then valid when others do it (papers C).
My conclusions are:
1) Climatologists earn much of their grant money by copying and pasting
2) Climatologists like to hold others to higher standards than they themselves maintain
3) The hockey stick is based on false statistics and scant data (i.e. copying text is an irrelevance to this fact)
Papers A and B (where Bradley copies)
Original:
http://www.ambiente.gob.ec/userfiles/2092/file/Cambio%20Climatico/Adaptacion/CLIMATE%20CHANGE%20IN%20THE%20TROPICAL%20ANDES%20PART%201.pdf (2007):
“In the arid and semiarid regions of the tropics and subtropics more than 80% of…” etc.
Derivative work:
http://www.geo.umass.edu/climate/papers2/vuille2008.pdf (2008) by Vuille, Bradley and others:
“In the arid and semiarid regions of the tropics and subtropics more than 80% of …” etc.
Papers C (where Bradley is copied)
Bradley, 1999 – as quoted by Mashey “Many natural systems are dependent on climate; where evidence of such systems in the past still exists, it may be possible to derive paleoclimatic information from them.”
http://books.google.com/books?id=SEO_RyNDJ0gC&lpg=PR1&pg=PA49#v=onepage&q&f=false
Chen, Xing, 2008 “Many natural systems are dependent on climate; where evidence of such systems in the past still exists, it may be possible to derive paleoclimatic information from them.” (etc.)
(i.e. this is a direct copy)
As an aside, one has to wonder – how much of the mountain of publicly funded climate “research” is cut-and-paste word-processor magic. Perhaps the Cuccinelli probe should be expanded. This may be an actual case of ‘It is worse than we thought’.
[Snip. Calling others here “deniers” gets your post snipped. ~dbs, mod.]
ZT, you somehow managed to elide the citation in the Bradley paper. The single sentence constitutes fair use. What happened with Wegman is that he copied huge chunks of text from Bradley w/o citing the source of the text.
Elise, tell me more. Who do you think has actually used “short-centred principal components? And when?
Actually, the wide circulation of Montford’s eccentric phrase “short-centred” is a giveaway. A lot of “let’s talk about the science” is just parroting.
But OK – Mann used decentred normalisation in 1998, probably by mistake.
Ugh. Sorry this is off topic (delete if it is too far off topic) but Nick Stokes’ comment above is just plain ignorant and wrong headed. It isn’t just Montford’s phrase and it isn’t eccentric – it is exacting and accurate, unlike Nick’s vague and inaccurate comment.
Nick tries to imply there are just two types of processing – centred and decentred PCA. The reality is that there are a multitude of conventions which have different behaviours and characteristics. If you want any hope of understanding the issues, you need to understand the properties of all the different methods.
For example, for PCA on data matrices (as MBH98 used), you can have column-centred PCA, row-centred PCA and double-centred PCA (meaning both row- and column-centred). All of these have different properties.
Furthermore, you can use the sample mean to centre (simply zeroing the sum of the column or row) or the population mean. The sample mean is most common because doing PCA is usually a statistical exercise (i.e., you don’t know the underlying characteristics and you are estimating them from the data) but in principle, the population mean would be more accurate, leading to (slightly) more accurate estimates of variance explained by the analysis.
There are different types of decentred PCA as well, such as unit-centred PCA – for example, the results you get from using a specific unit. This may be beneficial as the measurements themselves may yield a more accurate reference than the sample mean from the data.
Note that one of the possible reasons for decentring is that you may have a separate reference which is more accurate than the sample mean estimate – such as an external reference (when using units), or if you have more data that you are not including in the PCA. What Mann did was use a subsample of the data to centre. This method is unheard of in PCA literature. This is a *less accurate* estimate of the mean than using the full sample set, or a known reference. Why would anyone want a less accurate estimate? The inaccurate estimate directly injects an error into the “variance explained” calculations. Steve McIntyre, I believe, was the first to coin the term “short-segment” centred PCA, which is an entirely appropriate description of the flawed technique.
As we see, just lumping all forms of PCA into “centred” and “decentred” leads to misunderstanding. Different forms of decentred PCA have different properties, and it is highly relevant to specifically identify the type of decentred PCA being used. “Short segment” centred PCA is a form of centring which injects additional error terms into the analysis.
As for Wahl and Ammann showing that you can still “get” a hockey stick with conventional centring, the Wegman report clearly explains why this is flawed. They do so by tuning the results – changing the retained PCs, something which is carried out subjectively by the authors. Since tweaking minor aspects of the methodology can yield a wide variety of results (as explained in the paper “are multi-proxy reconstructions robust”, Burger and Cubasch 2005) which means that by tweaking these parameters you can get any answer you want – all Wahl and Ammann have to do is change things until they get the answer they want, then come up with a plausible justification for that decision.
As the Wegman report correctly observes, using this type of analysis is simply bad science. You can provide evidence for anything on this basis. Nick Stokes seems to be happy with this. Yuck.
“Rattus Norvegicus says:
October 11, 2010 at 12:04 pm
ZT, you somehow managed to elide the citation in the Bradley paper.”
Sorry, no – here are the two pieces of text. The later piece of text (by Bradley among others) does not cite the earlier piece of text. Please feel free to check for yourself:
http://www.geo.umass.edu/climate/papers2/vuille2008.pdf
http://www.ambiente.gob.ec/userfiles/2092/file/Cambio%20Climatico/Adaptacion/CLIMATE%20CHANGE%20IN%20THE%20TROPICAL%20ANDES%20PART%201.pdf
Vuille 2007:
“In the arid and semiarid regions of the tropics and subtropics more than 80% of
the freshwater supply originates in mountain regions, affecting more than half of the
earth’s population (Messerli, 2001). Much of this water is initially stored as ice in
mountain glaciers and then gradually released over time. Mountain glaciers, such as those found in the tropical Andes, therefore act as a critical buffer against highly seasonal precipitation and provide water at times when rainfall is low or even absent. At the same time these glaciers are particularly sensitive to climate change because they are constantly close to melting conditions. They are arguably the most visible indicator of climate change, due to their fast response time, their sensitivity to climate variations and the clear visibility of their reaction (glacier growth or shrinkage) to the public.”
Mathias Vuille, Bernard Francou, Patrick Wagnon, Irmgard Juen, Georg Kaser, Bryan G. Mark, Raymond S. Bradley (2008):
“In the arid and semiarid regions of the tropics and subtropics more
than 80% of the freshwater supply originates in mountain regions,
affecting populations downstream (Messerli, 2001). Much of this water
is initially stored as ice in mountain glaciers and then gradually released
over time.More than 99% of all tropical glaciers are located in the Andes
(Kaser, 1999) and Andean countries, such as Bolivia or Peru, rely to a
great extent on freshwater from glaciated basins during the dry season.
Mountain glaciers, such as those found in the tropical Andes, therefore
act as a critical buffer against highly seasonal precipitation and provide
water for domestic, agricultural or industrial use at timeswhen rainfall is
low or even absent. At the same time these glaciers are particularly
sensitive to climate change because they are constantly close tomelting
conditions. They are arguably the most visible indicator of climate
change, due to their fast response time, their sensitivity to climate
variations and the clear visibility of their reaction (glacier growth or
shrinkage) to the public.”
Just another example of a cut-and-paste scare story in the original peer-reviewed “scientific” literature of climatology!
My conclusion remains:
1) Climatologists earn much of their grant money by copying and pasting
2) Climatologists like to hold others to higher standards than they themselves maintain
3) The hockey stick is based on false statistics and scant data (i.e. copying text is a irrelevant to this fact)
Rattus Norvegicus says:
“What happened with Wegman is that he copied huge chunks of text from Bradley w/o citing the source of the text.”
Some questions, if you don’t mind:
1. Was any putative copying claimed to have been done illegally? Or was it simply against convention?
2. Produce evidence to support your unequivocal statement that Prof Wegman personally “copied huge chunks of text from Bradley w/o citing the source of the text.”
3. What would be the Wegman et al. motive for failing to attribute a source? You need a motive, and it must be more than just your opinion.
Explain why Wegman et al. are being held to such a high standard over what was probably either an inadvertent omission, or the work of one individual — when at the same time Michael Mann gets a free pass from you when his own written words showed him to be conspiring with Phil Jones to thwart the law. And for getting a colleague fired for not toeing Mann’s CAGW line. And for threatening to blackball journals that didn’t fix the game Mann’s way. And so on.
Even in the extremely unlikely event that Prof Wegman deliberately refused to attribute passages to Bradley, which act is more serious? That, or getting someone fired over different scientific views? Or conspiring to evade the law? Why is Mann still getting a free pass on those much more serious issues?
Take your time, I’m retired. I have all day.
Steven Mosher,
plagiarism is an act… quite clearly defined… you’re confusing whether you decide to prosecute that action… a moral decision that involves balancing consequences…
o’really?… your case to make
Rattus Norvegicus says:
October 11, 2010 at 5:43 am
“According to an update of the USA Today article, GMU has begun a formal investigation. Pass the popcorn.
For all of you who say this is not serious, well GMU seems to think it is.”
GMU will probably want to be careful with that. All formal inquiries, be it the courts or the legislature usually have mechanisms in place to ensure that evidence is submitted freely to them without external pressures being applied. This includes freedom from all forms of sanction for that testimony.
Otherwise, the Mafia could have hefty fines for breach of contract in place for all “employees” who break the code of silence and testify against the Don.
If GMU wish to punish Wegman for something he said to Congress, then they should beware of Congress turning around and ripping them a new one.
Lazar says:
October 11, 2010 at 2:06 pm (Edit)
Steven Mosher,
“no silly. as somebody who actually had to make the decision whether to bring cases forward and ruin a kid, I am saying that if one of my students did this i would not consider it to be plagiarism.”
plagiarism is an act… quite clearly defined… you’re confusing whether you decide to prosecute that action… a moral decision that involves balancing consequences…
#########################
quite precisely. In the same way that I think the virginia AG, while well within his bounds to act, is obviously not showing good judgement. I am not confusing anything. You are thinking that the bright line of the code is dispositive. It’s not. If it were then charges could be brought by machine and cases settled by algorithm. I would consider several things: 1. The venue. A presentation before congress has protect under the speech and debate clause. 2. The actor: Wegman is the lead author, and I would have to determine who wrote the sections in question. For me that would be the difference between committing plagiarism and failing to catch the plagiarism of others. 3. I would consider whether the material was presented as original thought per the assignment given. 4. I would consider wether the offended material was central to the assignment or merely prefatory or ancillairy ( like copying a bibliography replete with mistakes.. caught a student doing that once) 5. I would consider the gain/harm in the case.
Shoot-the-messenger-gate
Trip-wire-gate
What is auditing all about?
(1) Clarify the material, make sure you understand it. YOU ARE NOT EXPECTED TO BE AN EXPERT, you’re simply auditing the books. So of course you’ll need to explain the concepts, using standard texts but if need arises, modifying the language to make sure other auditors can understand – the key concept of the checking being replicability.
(2) Only when you’ve grasped the concepts can you do the auditing. Checking replicability, checking the fair use of data, checking the attainment of statistically significant results, is your area of expertise.
How can there be plagiarism in anything but an academic-nonsense sense, in rewriting standard text purely for the purpose of clarifying the issues? REAL PLAGIARISM IS ABOUT TAKING CREDIT FOR ANOTHER’S IDEAS.
Steven Mosher,
Someone stole something from me, I knew who it was, and I didn’t have the heart to press charges. It was still an act of theft. The people whose work has been copied without attribution… still have their work copied without attribution. It’s still plagiarism.
Wiki is a minor part, most is from two books, de Nooy (2005) and Wasserman (1994), see Mashey’s report pp. 119-128. I count around thirty paragraphs and over 1,700 words copied verbatim, one or two words changed here and there, not closely paraphrased. Looks like laziness. The stuff copied isn’t ground-breaking, mostly entry level definitions.
Steven Mosher,
I posted the previous comment before reading your latest… I’ve no problem with what you’re saying.
ZT says: “Plagiarism experts – how does the following look?…”
Like hypocrisy is alive and well in CO2-Cuckooland.
Did Mann provide references to Luterbacher et al?
ZT,
So now you are claiming that an author can plagiarize his own work?
” Rattus Norvegicus says:
October 11, 2010 at 6:45 pm
ZT,
So now you are claiming that an author can plagiarize his own work?”
I am not claiming anything. The absolute fact is that R.S. Bradley (and coauthors) recycled text from an earlier paper to produce a later paper.
Facts are likely a new concept in climatology – they are interesting – I recommend that you consider them more often.
My conclusions remain:
1) Climatologists earn much of their grant money by copying and pasting
2) Climatologists like to hold others to higher standards than they themselves maintain
3) The hockey stick is based on false statistics and scant data (i.e. copying text is a irrelevant to this fact)