Bishop Hill More on fudge and fraud
May 16, 2010 Climate: RP Jnr RP Jnr says I've misrepresented his views in the post before last. If so, then I apologise.
I'm still not sure that I understand Roger's views precisely. I think the confusion may be based in the semantics of the terms "fudge" and "fraud" and I want to explore the subject again here.
The original Spiegel article said this:
But what appeared at first glance to be fraud was actually merely a face-saving fudge: Tree-ring data indicates no global warming since the mid-20th century, and therefore contradicts the temperature measurements. The clearly erroneous tree data was thus corrected by the so-called "trick" with the temperature graphs.
The "trick" was of course, to truncate the divergent data, to replace it with the instrumental records for the same period and then to smooth the spliced series so that the join was no longer visible. The sentence is the Spiegel article seems to suggest that this "swap, splice and smooth" process could reasonably be described as a "mere" fudge.
This one sentence raises many objections. Is fudge actually distinct from fraud? Is fudging a trifling thing that can reasonably be tossed aside by attaching it to the word "mere"? And where does the "swap, splice and smooth" technique really fit in among these terms.
My guess is that lay commenters like me have a mental picture of scientific misconduct that encapsulates a whole bunch of transgressions that academics like Roger would treat as distinct. I think I'm right in saying that Roger sees fudging as a lesser transgression, although he is clear that he doesn't approve of this kind of thing. But he also says that in academia, hiding uncertainty is not generally considered as research misconduct. Again, I don't imply that Roger thinks this is acceptable - just the way things are. So if I have it right, in academia fraud and research misconduct are filed under "serious", and fudging and hiding uncertainty are in the "less serious" drawer. I'm just not sure this is how the general public see it. We expect full, plain and true disclosure. Hiding the extent to which you don't know something just gets put in the great big bin called "wrong". And making elaborate steps to hide the fact (swap, splice and smooth) just makes it worse. It certainly dictates against the use of the word "mere", as Spiegel did.
In the comments at Roger's thread, there seems to be much agreement that standards regarding disclosure of uncertainties are rather low in academia but much higher in the commercial world. This raises the intriguing question of how we should judge academic scientists whose work is impinging upon the real world. Should we expect them to have applied "real world" standards of transparency to their own findings? Or must we accept that what will be reported to policymakers by academics will be rife with deception.
Reader Comments (84)
Having read RP's latest, I am a bit more sympathetic to his point of view. What he is basically arguing is that academic fraud has a well defined meaning. It happens when you invent experimental results, destroy data, that kind of thing. The classic UK case was perhaps that of Cyril Burt, who reported on studies of twin pairs who do not seem to have existed.
Bad conduct on the other hand, is a much wider term. So what they have done with hiding the decline is presentational. No effort was made to deny that the data existed. There was a misleading presentation of it. This was bad behavior. But it was not academic fraud.
Academic fraud would have been had the decline data been altered in the original article publication.
A much harder question about the HS is the use of PCA. Maybe the harder aspect of the general question is about the use of statistics, and accepted versus home grown methods. If you perform what you describe in the original article as the use of PCA to extract the factors, but when computing it, you use a novel method previously not recognized as being the computation of PCA, and do it without explanation or justification of the method, is that academic fraud? Or is it simple incompetence and bad practice?
Here perhaps the drug approval example has a bearing. Suppose someone had done 'short centered' PCA on his experimental results, suppose in consequence he had extracted what later turned out to be a wholly illusory factor, and suppose that treatments had been based on this study, with some avoidable deaths resulting. What would we call this?
Michel
Yes, it does seem to be down to the definition of fraud as used in academic circles. The question that I'm asking myself is whether this is the definition that was used in the original Spiegel article, which was directed at lay readers and not academics.
I'm also bothered by the terms "merely" and "fudge", used in close proximity. I read this as brushing the criticisms aside, which is wrong.
Tim - afraid I agree with Barry here. Mann et al appear to have had ample opportunity to admit sloppy science but have opted to conceal their methodology and their data and chosen instead to defend their stance over many years. That rather rather transforms the charge against them. It was not sloppy - it was deliberate.
I vaguely recall a suggestion of a petition some time back signed by AGW believers with the question about the number of PhD signatories and the remark about it is only the numbers that count nobody will question the qualifications. I think that fits nicely with the attitude of misleading presentation and bad behaviour that Michel posts above.
It is the ability to mislead, trickery, with a cover story to fall back on if and when they are rumbled. Is that fraud or bad conduct?
I can feel a new YouTube song coming on: "Who Hid the Fudge?"
Two points, I am one of those who would help Mann if he is charged on the basis of a fishing expedition by the AG of Virginia. My reasons aren't entirely altruistic, although I do take into account the push by AGWers to have dissension suppressed, and dissenters jailed, but the reality is that the stakes are so high if the AG doesn't find anything then Mann becomes Galileo, and Galileo he ain't. Some are already comparing him with Galileo.
I believe he manipulated the data to get his "hockeystick" by the judicious use of proxies and hitherto unknown statistical methods, but we've just seen the greenwashes of Jones on a set of evidence that would have anybody put, as the police say, "bang to rights". In other words evidence so incontrovertible that there was no chance of being found innocent. But they did, right there with the whole world watching, the inquiries excused the most egregious, unscientific behaviour and cleared him almost everything except for being a little sloppy. We need to tide to turn a lot more than it has politically before we can put the authorities in a position to do anything about malfeasance in the climate science community. Vis. Gavin Schmidtt's shameless labelling of these excuses for scientists as "blameless". Time, I'm afraid is the only thing that will allow this abuse of the science to be revealed, too many people are true believers, and too many scientists have stayed quiet, who will be disgraced with the charlatans for doing so, for us to move to a rapid denouement of this situation.
I repeat though, I don't approve of political fishing, and would do my utmost to protect any scientist from it. What we want from science is the truth as it is known, if we put political pressure on scientists, even those as insignificant as Mann, we will get the science the politicians want to see, and no one can gain from that.
Michel,
Again we are back to these specious claims that academics are somehow different to the rest of the human race, and are a somehow privileged group, with their own legal code, the hem of their robes withdrawn from the populace.
There seems to be a strong feeling amongst non academics, voiced in various blog and newspaper comments, that this thought is erroneous, and may be firmly placed where the sun don't shine.
To claim that there is this special 'academic fraud' that is well defined and somehow merits a slap on the wrist rather than obedience to the law, is nonsense as far as normal citizens are concerned, and they want none of it.
If, and I repeat if, someone e.g. an academic, has committed fraud, then they are subject to the same censure and penalties as anyone else. Whether they are in fact guilty of the charge is for a court of law to decide.
There appears to be strong support from the blog reading general public that this should occur, or so it seems to me reading the comments.
"He (Mann) is not a criminal and no good will come from trying to make him one."
Hold on.
The Virginia A.G. seems to be investigating possible financial misdeeds relating to research grants. If Mann were convincingly convicted of a financial crime by a jury of his peers, I think that some people who currently believe that "the science is solid" would begin to have doubts.
Tim & Michel
I too take the hard view and agree with Chuckles, Martin A., and others. Mann et al have cost the tax payer of the world a great deal. And I am still waiting on Rodger Dodger to answer my question about what is his opinion of this brutal fact.
As Ed Forbes noted:
"In this country, it is wise to kill an admiral from time to time to encourage the others." [Voltaire ]
The Chinese also have a saying:
Chop the head off a chicken to scare the monkeys
Wisdom of the ages, it is.
Interesting, I did a fudge in a computer model recently because the answer - a temperature reading in a pasteurizer was 'obviously' wrong i.e. it was 2 or 3 degrees down, despite the inputs being otherwise correct. But it was difficult to simulate secondary heating actions correctly in the heat exchange without getting very complex. Using a linear equation to calculate and add the missing factor worked fine. Two wrongs do occasionally make a right. However, in this case, I knew what the answer should be whereas in the climate models it appears to be more of a case of I know what I want the answers to be - a subtle but deadly difference.
I don't know whether the subsequent and other conduct amounted to criminal fraud in the eyes of the law. I think the behavior in 'hiding the decline' was very bad and unacceptable.
I don't think anything amounting to fraud on the Cyril Burt lines has been shown. With the possible exception of some of the statistical treatments. One would need to show that the tests were changed and the methods deliberately doctored. I don't know if that has ever been proved, but running deliberately misapplied stats on unaltered data is getting close to Burtian fraud. If that is what was done, though, I don't think it has been proven.
We need to be able to express our disapproval in terms other than those of fraud. It does not mean that the conduct is acceptable, just because it does not fall within the term of academic or criminal fraud.
Obviously the amount of money involved with both the grantee and the grantor needs to be factored in.
Michel
Just what is your definition of Fraud? You are an intelligent person and I do find your posts interesting. However I am confused by your position on this.
Fraud is not assigned a category it is not acceptable and unacceptable fraud, there is no such thing as academic fraud. Fraud is what it is …..deceit, trickery, sharp practice, perpetrated for profit or to gain some unfair or dishonest advantage. Removing the divergence data wasn’t done to present a nice tidy story it was done because the divergence could not be explained and they attempted to avoid awkward questions. It was deceit, trickery, sharp practice, cheating, underhand or in a word, fraudulent.
Climate change exists in the silicon chips of supercomputers
-Stephen Schneider
Global warming has lost its standing as a celebrity social problem...and it now receives intermittent media attention that is as often negative as positive. The issue will not regain its celebrity standing...simply by an intensification of claims making by concerned activists. Another spike of attention will require novelty and drama. In effect, potential agenda-setters will need something more to sell - or piggyback on - if they are to bring widespread attention and concern to the issue
-Sheldon Ungar, Social scares and global warming: Beyond the Rio convention, 1992
The Hockey Stick is the bridge by which we all crossed over from the placid island of climate science into catastrophic alarmism. Who's going to forget about it in a wink and a nudge?
martyn
While I agree with you, I would like to hear Michel's position. I respect him just as I respect you, but he is seeing something I am not. I would like to understand it. We can all learn from one another even if we disagree.
I suspect Michel agrees with us in principal but not in the use of the word. It may be merely a semantic issue, or perhaps something else. I am curious, that is all.
What is really interesting is the fact that the Spiegel journalist buys in the upside down logics of our climate scientists :
I use a proxy to reconstruct a sery from the past till 1960 or 1980, and when more recent proxy data don't fit with the reconstruction and the model I strongly believe in, it only possibly means that my recent tree data measurements are obviously wrong ; consequently I should not mention unhelpful data which could limit the public faith in my tree ring model, and due to the lack of alternative, as a most needed solution, I rightly insert in the curve actual temperature data which are totally inconsistent from a methodological point of view.
Just an upside down apple pie without its flavour !
Not so far from an implicit yes to the basic receipies of propaganda.
Having read all the debate on this thread and RP Jnr blog as well ... what odds on the "right" outcome when the emails inquiry finally reports? Preliminary conclusions are due to be reported within the next couple of weeks (before the end of May).
I gather they have asked for access to all the unleaked emails from the server to assist their analysis - they have also recruited someone expert in peer review ... so presumably they intend to produce more than a 5 page report?
My guess is that they will not want to criticise any individual at all (fear of libel?).
Anyone want to hazard a guess as to the likely possible outcomes?
I hear on the grapevine that the report is more likely to be seen in June than May.
Is that a grapevine that has tendrils particularly close to the enquiry?
Because on their workplan page they state "The team expect to have at least preliminary conclusions by Spring 2010" while on the FAQ page they state "When will the Independent Review submit its report? The University of East Anglia has asked the Review team to submit its report in Spring 2010. The Review is aiming to meet that target, but precise timing will obviously depend on the work itself." They also helpfully add "What does 'in the Spring' mean? Appropriately, meteorologists define Spring as March to May. We will move as quickly as we can, but we can’t give a precise date now."
So rather a lot of people will be expecting preliminary findings before the end of the month ... and they have given no indication to expect otherwise.
My source said in late April that "they've got a lot more work to do" and "it will still be quite some time before they can report".
"Or must we accept that what will be reported to policymakers by academics will be rife with deception"
Everyone is selling something. Just because someone has a PHD doesn't change their ethical standards.
Is it a delay that grows longer and longer and when everybody forgets about it, so do they?
That is a possibility. Any insights?
As I read the responses here I get the distinct impression academics are given considerable license in the general view -- academic freedom of inquiry I get, appreciate and support. The big "however" is that I don't see this translating to loose standards in the reporting of those inqiries.
Why shouldn't science be held to higher standards of observation and reportage than the businessman, for example? Any businessman caught presenting a prospectus as dodgy as Mann's proxies or the WMO cover graphic would be subject to intense discovery and prosecution to the fullest extent of the law. Some financial frauds have cost biillions but catastrophic global warming panic has a price tag in the trillions -- the stakes are immense and it is incumbent on researchers to deliver all caveats and conclusions no matter how "untidy". We are not talking about ripping off a paltry few million investors greedily seeking implausible returns from funds they apparently did not immediately need but the suppression of rising living standards of developing peoples, re-engineering the world's energy supply and restructuring society.
Can we excuse lesser standards because "they are only academics"?
Well, they are only fudging the world economy, society's energy supply and people's standard of living and they are only doing so for the small personal gain of career advancement and pecuniary return so we really shouldn't investigate too closely...
Really?
Perhaps I missed it, but I haven't seen a pointer to Josh's comments on Roger Dodger.
HERE
I might also point to the link under the cartoon. Very interesting reading which includes the following:
"Those two sentences have resulted in several lengthy comment threads, with many comments critically directed at me, fueled by a cheap-shot post by Andrew Montford who has since apologized and asked for clarification, which is fair enough. Here is the clarification."
Jezz, I thought is was a shot gold shot, Bishop.
I think there are two things: fraud as defined in the criminal law of the jurisdiction. This covers stuff like selling a property to which one does not have title, willfully filing false accounts, claiming child benefit for non-existent children. Claiming expense reimbursement for non-existent research assistants or office expenses or travel expenses. This is a well defined and legal issue, one refers to statute and case law, and it is prosecuted in the UK by the CPS.
There is then a different thing, which is not to my knowledge prosecuted as fraud by the CPS, which happens in academic life, and that is usually referred to as academic fraud, with the specific meaning that one has misrepresented, lied about, or invented experimental results. This is what Burt seems to have done, and there have been recent cases in the cloning area where it seems to have been done.
It is usually distinguished from simple research incompetence by the fact that incompetence is due to good faith if idiotic errors, whereas academic fraud has the element of knowledge and intent, as well as the element of falsification.
Academic fraud may also be criminal fraud, if its engaged in in the right circumstances, but it is often not regarded as such, not prosecuted, and when proven, the usual penalty is confined to dismissal.
I do not think Mann and Jones engaged in behavior which is fraudulent in the legal sense, and exposes them to prosecution if in the UK. I haven't seen any pro forma case to argue this.
Did they engage in academic fraud? Again, I have seen no evidence that they deliberately falsified data or results. They did not make up Briffa's data. In a presentation, they did not show all of it, in a way that was misleading. That still is not what we usually call academic fraud. It is bad or misleading representation of results, its thoroughly bad conduct, but its not the wilful falsification of data.
What about the statistics, then? Here we are in a very grey area, and I think it is possible to commit academic fraud in general terms in the area of statistical manipulation and representation of the results of statistical tests, though I know of no evidence to point to it having been done in the specific case under discussion.
The kind of thing that might count as academic fraud in this area might be the following example. I deliberately do a t test with a subset of the data, and then report the results for significance, as if I had done the test correctly using all the data.
Was that academic fraud, to represent my results as being those of having used a t test to assess significance? Maybe. Probably, in this particular instance, involving something as simple and well known as a t test. Once you have ruled out incompetence in how to do t tests, the remaining explanation is wilful. But it is hard to prove this. And in more complex areas of stats, it will sometimes simply be a difference of opinion about what tests are appropriate, and how to do them.
In general, I think we should not use the expression 'fraud' as a means of condemning behavior. We should be precise about what has been done, and why it is objectionable, and not use words which can be used with precision, simply on account of their emotional connotations. Someone with a good understanding of statistics who has deliberately done what I described above has gone beyond incompetence and misrepresentation in a context of presentation of implications, and come very close, maybe over, the borderline where the results are actually being falsified. But calling it fraud does not really add much to a description of what was done, and why it was wrong, which will be much preferable to calling names. As well as more effective in combating the errors.
Michel, you do know about Mann's 'censored file' with respect to MBH '98, don't you? If not, please review the matter at climateaudit.org
Apparently, Steve McIntyre is not convinced that this is evidence of fraud. I believe it is evidence of fraud, but not proof of it. Intent is so very necessary for the application of blame and on this matter we remain ignorant.
But inquiring minds want to know.
=====================
Michel
While I see your distinction, and indeed agree that there has been a good deal of "academic" fraud over the years -- for example Piltdown man -- I believe you are ignoring the fact that what Mann et al have been doing is criminal in that they are both using the fraud to finance their activities with grants given based on their fraud (how is that different than stuff like selling a property to which one does not have title, willfully filing false accounts, claiming child benefit for non-existent children?). And far worse costing us all trillions in unwarranted taxes. In the later case, they have clearly entered into a quid pro quo with those who control those grands in exchange for "evidence" to support their levying those taxes as well as profiting from the "Carbon Trade".
I agree with Kim that there is clear evidence of criminal fraud, but not sufficient proof for conviction at least yet.
I appreciate your response. Thank you.
Yes, I do know about the censored file. Look, I could have folders called 'censored' on my computer. Anyone could.
The question is, what exactly is the conduct which is being classed as the commission of a fraud in criminal law, or academic fraud in the ordinary sense? Mann would say, I had trouble with the data, I felt strongly there was a trend in there that I was not getting at. I did quite a few runs, and when I was not satisfied, I moved the runs into an archive folder. It was unfortunately titled, but that's what I did, I kept them in case I needed to return to them later, and anyway, they were part of the study, I felt obliged to retain all results even if they did not show what I thought was the real situation.
I don't think it is useful to describe what he did as the commission of a fraud. He was wrong of course in his treatment of the data, the whole thing was a total cockup. But that is different. Not everything that academics do which is either mistaken, stupid, or both is helpfully classified as a sort of fraud.
Take a case where there might be academic fraud, if the facts are as alleged. The Jones Chinese study. Suppose that his co-worker really did misrepresent the existence of the meta data, and said it was there when it wasn't. I think this is the gist of Doug Keenan's complaint. To state that your data has a certain provenance, when you know it does not, would normally count as academic fraud in the usual sense of the term. I don't know whether Keenan's complaint is correct, but if it is, if the facts are like that, it would be a case of academic fraud. But not by Jones of course. The worst that Jones could be guilty of is failure to check the sources of a paper of which he is the co-author. Not good. But not fraud. For it to be Jones' fraud, he would have had to know that the data was not as represented, and would have had to knowingly then agree to falsely representing it. I don't think he has ever been accused of that.
We need to use these terms much more specifically and precisely, and we need to focus on what is important. In the end, whether it was fraud, wishful thinking, stupidity, is not nearly so important as that the HS is disastrously wrong as a representation of the probable course of past and present temperatures. This fact, and the why of its wrongness, are what counts, and to spend too much time about whether what the authors did can be called fraud does not really advance public knowledge of that.
In the end, whether it was fraud, wishful thinking, stupidity, is not nearly so important as that the HS is disastrously wrong as a representation of the probable course of past and present temperatures. This fact, and the why of its wrongness, are what counts ... "
In that case it would be tremendously helpful to all sides if we had an admission of that - even at this late stage - would it not? But we haven't. It seems that "they" are intent on maintaining what everybody else perceives as a deception.
Main Entry: fraud
Pronunciation: \ˈfrȯd\
Function: noun
Etymology: Middle English fraude, from Anglo-French, from Latin fraud-, fraus
Date: 14th century
1 a : deceit, trickery; specifically : intentional perversion of truth in order to induce another to part with something of value or to surrender a legal right b : an act of deceiving or misrepresenting : trick
2 a : a person who is not what he or she pretends to be : impostor; also : one who defrauds : cheat b : one that is not what it seems or is represented to be
synonyms see deception, imposture
"In that case it would be tremendously helpful to all sides if we had an admission of that - even at this late stage - would it not?"
Yes indeed. One of the silliest and most counterproductive tactics of the AGW movement is its determination to defend the Hockey Stick, including those dodgy statistics, to the bitter end. They must have converted more people to skepticism by that tactic than any number Exxons ever could have done.
Dear Bishop
RPJr has tripped up - just a bit. This is not 'getting back' or anything. just a humble reminder that anyone of us can screw up.
http://nigguraths.wordpress.com/2010/05/19/climate-fudge-fraud/
Been re-reading the minutes of the Email Review (available up to 1 April)
Meeting of 20 March
GB gave an oral presentation on the submissions received in relation to the remit and issues identified by the Review. [i.e. The Review are identifying additional issues of their own?]
DE presented an analysis ... of the emails which were the subject of unauthorised release from the UEA. [i.e. no longer an assumed hack from outside?]
The Review agreed that the issue of peer review is a very important consideration. [i.e. not being swept under the carpet then?]
... should the Review find matters of immediate and pressing concern, then it clearly has a duty to inform the University. Natural justice demands that both the UEA and members of CRU should be told of any critical findings which are to be leveled at them, and be given the opportunity to answer. [i.e. there is at least some possibility that there may be matters uncovered that may be critical and of immediate and pressing concern?]
Meeting of 1 April
The Review noted the actions being taken as regards searching the entire CRU backup server record. [i.e. should one assume there is clear justification for extending the investigation to areas as yet unseen by the public?]
Continuing analysis of the submissions was enabling the matters identified in the initial issues paper to be focussed more sharply, while it was clear that the essential questions for examination had not changed. [i.e. the original questions have presumably not yet been discounted]
(Preliminary report still expected this month.)