About these ads

Retraction Watch

Tracking retractions as a window into the scientific process

Brian Deer’s modest proposal for post-publication peer review

with 50 comments

brian-deer-d-fullBrian Deer’s name will no doubt be familiar to many Retraction Watch readers. Deer, of course, is the award-winning investigative reporter known for his reporting on numerous medical issues, including Andrew Wakefield’s now-retracted research into autism and vaccines.

Deer is giving a talk next week at the UK’s “Evidence Live” conference,and has a proposal that he hopes will make it more difficult for dishonest researchers to hide their misdeeds — and make it easier for journals to retract fraudulent papers. He has expressed concern before that voluntary codes have no teeth. Deer is proposing an amendment to the ICMJE’s Uniform Requirements for the Submission of Manuscripts to Biomedical Journals:

Post-publication review

Papers may be subject to post-publication editorial review and retraction.

A failure by authors to disclose and speak to the project’s full methods and data, in recorded editorial interview, or upon request by an appropriate panel, employer, funding body or court, shall be grounds for retraction.

Deer welcomes feedback from Retraction Watch readers, so we were glad to have the opportunity to present his proposal for discussion here.

About these ads

Written by ivanoransky

March 22, 2013 at 9:09 am

Posted in Uncategorized

50 Responses

Subscribe to comments with RSS.

  1. The current guidelines are adequate – it is just the journals don’t abide by them. So why would they be interesting in paying anything more than lip-service to Deer’s proposal?
    Or as Einstein put it: “Insanity: doing the same thing over and over again and expecting different results.”

    littlegreyrabbit

    March 22, 2013 at 9:27 am

    • I agree.

      Average PI

      March 23, 2013 at 6:07 am

    • Einstein didn’t actually say that. It’s also not wisdom: “I applied for three jobs and didn’t get any; it would be insane to apply for another one and expect different results.”

      Frank

      March 24, 2013 at 9:24 am

  2. I would put some sort of statute of limitations on this. Following these guidelines strictly, you could probably get most papers published before the advent of electronic storage retracted. Particularly if the authors are no longer available to respond.

    The Iron Chemist

    March 22, 2013 at 9:37 am

  3. This will strike fear and awe into the hearts of science-fraudsters. If successful It is the beginning of the end of science-fraud.

    Deer is no doubt a brave man who will need our support, but I for one am fearful he will come up against many brickwalls. Brickwalls made by the very science-fraudsters his forward-thinking proposals threaten.

    As an extension of Deer’s mindset perhaps grant-awarding bodies could be restricted from awarding further monies to science-fraudsters or those reticent to show their present or past published raw data openly.

    stewart

    March 22, 2013 at 11:02 am

  4. Wrote about post-publication peer review and about Retraction Watch here:
    Peer-reviews don’t protect against plagiarism and wrticles retraction. Why?
    This is a strong argument for perpetual, open, peer-review.

    chorasimilarity

    March 22, 2013 at 12:27 pm

    • chorasimilarity, there is already “perpetual, open, peer-review”. In its traditional form it’s part of the normal fabric of science. In the electronic age it’s been extended to include the sort of investigative analysis of image manipulation and plagiarism that is found on blogs, including RetractionWatch. Some journals allow comments to papers that could constitute a form of post-publication peer review, but in my experience this tends towards the sort of unproductive opinionating, grandstanding, trolling and point-scoring found more generally on blogs.

      Plagiarism is a special case. There’s no reason why plagiarism should necessarily be identified in traditional peer-review (although there are some, perhaps many, cases where it is). Nowadays plagiarism should be identified by publishers using the standard software of the sort Universities use to assess student’s electronically-submitted essays and project work. It’s scandalous that this doesn’t occur right across the board. Plagiarism could be stamped-out overnight. We could discuss reasons why this isn’t happening…

      chris

      March 23, 2013 at 10:05 am

      • chris, re “in my experience this tends towards the sort of unproductive opinionating, grandstanding, trolling and point-scoring found more generally on blogs,” could you give me an example?

        chorasimilarity

        March 23, 2013 at 3:23 pm

        • In my experience on PubPeer.com this is absolutely not the case. The conversations that I’ve seen there (at least in neuroscience) have all been very productive.

          Paul Patrick

          March 29, 2013 at 5:16 pm

          • Paul Patrick, thanks for pointing to PubPeer.com. I look forward for similar approaches in mathematics. There seem to be something on the brink of appearance, based on selected-papers networks, see a comment by John Baez here.

            chorasimilarity

            April 1, 2013 at 5:53 am

  5. At this moment, there are three comments: The first, by littlegreyrabbit, is concerned that journals aren’t strict enough. The second, by The Iron Chemist, is worried that journals will be too strict. I’d say that this shows that Deer’s proposal is about right.

    The third comment, by stewart, is a bit zealous for my taste, but I agree overall.

    Ken Pimple

    March 22, 2013 at 12:54 pm

    • It is not that the journals aren’t strict enough, it is they have policies that are there entirely for show. This is a reasonably representative policy:
      “Scientific Misconduct

      Journal X will evaluate the credibility of all allegations of scientific misconduct, e.g., suspected fabrication or falsification of data, double publication, or plagiarism. If the Editor-in-Chief determines that an allegation has merit, s/he will first attempt to address the matter with the Corresponding Author. Should that fail to resolve the situation satisfactorily, the Editor-in-Chief will contact the institution of the Corresponding Author to request an investigation; the Editor-in-Chief may also contact the co-authors and/or the funder(s) of the published research.

      Until the matter is clarified, no additional submissions by any author of the disputed manuscript or published article will be considered for publication. If scientific misconduct is confirmed by the institution, and no request for retraction is made either by the institution or by any author, the Editor-in-Chief will report his/her findings to a representative of the Journal X Publications Committee, and in consultation with a representative of the Journal X Council, will decide appropriate action. ”

      This, I submit, is an excellent policy, it is just that Journal X – and I am sure most other journals – doesn’t adhere to it (or if you want exactitude – sets an impossibly high bar to determine if an allegation has merit). In general, journals don’t refer reported misconduct to institutions, and institutions don’t investigate.

      There is no point in making policies more ornate if they are unwilling to implement their already excellent existing policies.

      Besides which I tend to agree with Nigel Tufnel below – peer review does not always bring out the best in scientists as it stands

      littlegreyrabbit

      March 23, 2013 at 2:47 am

  6. The Iron Chemist is right about a statue of limitations. A statute of limitations already exists, because it is implicit in the currently accepted standards for data sharing. The problem is that these standards are rarely enforced. A recent example of the difficulties experienced by those trying to obtain raw data is provided here.
    With respect to current guidelines, though one can argue that they are sufficient, they are only guidelines and it may be convenient/easier to ignore them. One only has to look at the very different response of journals to the re-use of data across papers and the legions of dodgy western blots. In some instances the journals retract without hesitation. In others, corrections are allowed, on some occasions we even see corrections to the corrections. Thus, if one counts the number of papers with re-used and/or manipulated images and data in the Melendez oeuvre and subtract the number of retractions the answer is not even close to zero: it is an unreasonably large, positive integer. This problem was discussed extensively in Science Fraud and previously in Abnormal Science and has been repeatedly raised in comments on Retraction Watch.
    Teeth with a decently powerful jaw are needed and I hope that Brian Deer is successful in his enterprise.

    ferniglab

    March 22, 2013 at 1:26 pm

    • A little more on my idea:

      Editors would be able to offer a conference call to an author whose work has been seriously challenged, with a specialist on the line, to take down a recorded account of what the author says. This is not intended to be prejudicial, and, being in the professional standards statement, isn’t rude.

      I think some number of authors will decline to take part, not least because a digital record of their story could be admitted later if there was an institutional inquiry. If so, simple, paper retracted. No fault assumed. Maybe the author has a good answer – but it had better be the same answer they decide to rely on further down the road. This could spare the potentially enormous resources required (sometimes including retaliatory litigation) which I think cause a lot of journals to ignore problems. OK, there’s no blood on the carpet, but a job has been done.

      Institutions can save too. In serious cases, a lot of staff time and possibly litigation could be avoided if the proven offense is refusing to cooperate in a forensic editorial interview – the possibility of which was a condition of the paper being accepted. Again, this may not yield the whole story, which might be ventilated at any hearing, but the main thing is to get a result.

      I’ve come across a lot of scientists who say that reproducibility is all that’s necessary to safeguard the literature. But that presumes that the literature is the preserve of specialists who know the context within which fraudulent papers are situated. Increasingly, at least in some areas, it isn’t their preserve. All kinds of people wrench false papers from their context and present them on the internet as evidence of something – sometimes harming others with untrue medical claims and other wrongdoing. The literature needs cleaning up so that those folk who, say, run lists of papers for which they can’t even understand the titles, will find it much more difficult to mislead or be misled.

      I recall very vividly one editor telling me he had phoned an author who said the allegation wasn’t true, and another editor who brought four authors to his office, who likewise said the allegation wasn’t true. But no record was made, and their stories subsequently changed.

      So I’m thinking of an editor saying: “Oh, I’m sorry about this Dr Doe, but the uniform requirements say we have to do this. Don’t take it personally. I’m sure it’s a misunderstanding. Professor Jones is on the line,and I’ll put the recorder on now…”

      Brian Deer

      March 22, 2013 at 2:07 pm

      • That seems like a terrible idea. From your description it seems that a competitor in the field or someone with a political agenda that is not too happy with the implications of my research convinces an editor that I need to be “cold-called” to make an instaneous justification of my methods/analysis on tape. Apart from the fact that a recorded cold call is likely to be a very poor way of establishing the scientific rigour of my methods/analyses, it sounds like a dream scenario for the politicslly-motivated bully to attempt to browbeat scientists whose work they don’t like.

        Much more straightforward would be a better effort to ensure journals enforce the requirement that appropriate data/materials are accessible post-publication.

        I don’t understand what this means:

        “But that presumes that the literature is the preserve of specialists who know the context within which fraudulent papers are situated. Increasingly, at least in some areas, it isn’t their preserve.”

        chris

        March 22, 2013 at 3:30 pm

      • This will not work and i don’t think there will be buy in from journals. My recent experience in this matter indicates that there can be a substantial gap between theory and practice.

        As Mathias Brust argued in a recent article in Chemistry World entitled “Safeguarding science against fraud demands debate” we need debate. A closed discussion, whether on or off the record between authors and editor will serve no one. The editor is generally not a specialist in the technical aspects of the paper. In the case of the recent debate on striped nanoparticles, the process took 3 years, merely because the editor of Small actually followed procedure, whereas the authors whose work was being challenged played for time. The people capable of evaluating a paper are the peers. The appropriate peers may not have reviewed the article, but they are just a small sample of the peer group. So there is a need for post publication peer “review”. However, this doesn’t happen very often. The debate alluded to above is unusual because it is actually happening. There are plenty of sceptics in the field, but no one wants to raise their head above the parapet.
        given the “system”, which requires people to publish, then we need to consider the “teeth” I alluded to in my earlier comment. These might include:
        1. Requests for raw data used for figures in papers (including supplementary) to be honoured within 1 month for data in manuscripts published since the statutory time limit. Such requests to be cc’ed to journals.
        Request not honoured? Paper retracted.

        2. Journals to consider papers that challenge head on one of their previous publications solely on the grounds of technical merit. The “novelty” criterion cannot apply in this instance, since the original paper was deemed “novel”, its challenger is too. Witness the debacle over DNA with a sugar-arsenate backbone and the tremendous effort of Rosie Redfield and colleagues to refute the highly publicised claims for the existence of this alternative backbone.
        An “open” peer review of the new paper(s) and the previous one could then be initiated by the journal in a moderated (just to keep the debate clean and to prevent trolling) comments section.
        The editor might then produce a summary once the debate died down, which would be a justification for retraction or not of the original paper. The verdict of “not proven”, used inScottish law, may be suitable in some instances.

        3. A better mechanism for accountability for serial/multiple offenders at institutions, journals and funding agencies.

        ferniglab

        March 22, 2013 at 5:17 pm

      • Sorry, but only a nonpracticing scientist could propose something of this sort and think it has any chance of getting traction. Any practicing scientist know to well that there are plenty of entities out there who would love to start harassing scientists with false accusations.

        Average PI

        March 23, 2013 at 6:18 am

      • Just out of interest,how does this phone call scenario work out where the authors may have limited ability to speak English, particularly in a conversational sense?

        Putting a non native Engish langugage speaker “on the spot” seems to have great potential for misunderstandings, and a non scientific translator in the mix could cause even further problems.

        PWK

        March 25, 2013 at 12:37 am

  7. What do you think about this. It goes against the 5tyh amendment so I wonder what would happen at US institution (current policies do not say that failure to cooperate is ipso facto basis for retraction or judgment of wrong doing.

    Chaffin, Lajean

    March 22, 2013 at 3:46 pm

  8. This proposal sounds very good, except that I see it weird with the phone interview. This must be in the form of Q&A, let people breath and think and let it be recorded in writing. You can also go with it repeatedly, no problem.

    The best part is that authors MUST answer. And here, universities must subscribe to such policy. If the author does not respond, university must respond.

    The part about limitation period, I think, is wrong because if some fraud and especially – plagiarism is in libraries and e-available, this mere presence of the material is a violation of law, not just university rules, but – law. In plagiarism, falsification of the fact of authorship can be actionable in court 50 years after the author’s death.

    But, generally speaking, cleansing of literature, i. e. science, is the duty of the scientific institutions. I know, scientists have right to publish, but institutions must provide forum for discussing publications if, I repeat – if the author wants to send paper with the name of the institution. I think that it should not be automatically allowed. In this case, journals will immediately notice the absence of the name of the institution. May be this will work.

    I have started a web site with my own views on the fraud problem. I believe that… OK, if you are interested, please see it: http://sciencemishandled.wordpress.com/

    pyshnov

    March 22, 2013 at 4:12 pm

    • There needs to be a statute of limitations I think pyshnov (The Iron Chemists point). I would have though a period of somewhere in the period 5 – 8 years would be appropriate. One should be realistic about life and stuff.

      I was asked a couple of years ago for raw spectroscopic data from a paper I published around a decade ago; the editors were supporting the indexing of a spectroscopic database. I tried to hunt down the raw data but I simply couldn’t find it and had to apologize to the editor that I must have dumped it at some time in the past. The notion that my paper, cited numerous times subsequently, should be retracted as a result is absurd. Of course that’s not quite the same as the current expectation (in the electronic age) that data is more comprehensively archived.

      I would expect that it’s reasonable that some set of appropriate “raw data” is accessible for say 8 years post-publication. But at some point any particular publication becomes “embedded” in the scientific literature and if any meaningful flaws (that might require recourse to the raw data) haven’t been identified by then, it’s pretty likely that the field has moved on.

      The situation with your example of plagiarism is different. Plagiarism can be objectively identified explicitly with respect to written words in a paper in the context of previous written work. So plagiarism can be identified in a time-independent manner and it would be appropriate to retract plagiarised work even many decades post-publication and even if the authors are deceased.

      chris

      March 22, 2013 at 4:46 pm

      • Worse, had you found the data, they may have been in some unreadable software/hardware format!

        ferniglab

        March 22, 2013 at 5:19 pm

  9. Yes, Chris, you understand it the same way as I do: if there is as you say “time-independent” proof… But, in this logic, you don’t need any limitation period. And, this rule can be applied to cases other than plagiarism.

    pyshnov

    March 22, 2013 at 5:24 pm

  10. Reblogged this on Ferniglab's Blog and commented:
    An excellent discussion has started on this Retraction Watch post on Brian Deer’s proposals to make life more difficult for dishonest researchers.

    ferniglab

    March 22, 2013 at 5:26 pm

    • I mentioned somewhere earlier – even in the original publications, how about disclosing the names of the reviewers? I know some journals are doing this. If high profile journals do this – that would be a revolution, I guess.

      Ressci Integrity

      March 22, 2013 at 7:41 pm

      • Deer’s now has a hint of the resistance he will face. He must fight on.

        Disclosing the names of reviewers for articles would be entirely consistent with an open access policy.

        We could then develop a database which could be cross-referenced against a database of grant reviewers.

        The results will surely set us all free of previous, present and potential science-fraudsters thereby accelerating science, and accelerating the benefits of science to society.

        We would then do what the scientists amongst us do best – science!

        Stewart

        March 23, 2013 at 5:53 am

      • I don’t see how removing reviewer’s anonymity will help with the particularly serious instances of scientific fraud that we should be especially concerned with. In my opinion this will actually have a negative effect in two respects:

        (i) A paper is found post-review to be fraudulent. The reviewers who may well have been perfectly diligent are tainted undeservedly. They may decide that reviewing papers is more trouble than it’s worth….or editors might consider that it’s best not to use these “flawed” reviewers. Result: good reviewers lost from the system just when they’re sorely needed.

        (ii) A young early career scientist is asked to review a paper by one of the major figures in a field. Lacking anonymity s/he may be intimidated from writing the sort of review that would be particularly useful (in my experience young scientists make the best reviewers because they have a rather clear and progressive view of their chosen vocation). Result: less useful reviews.

        Non-anonymity could have a positive effect on one particular very modern type of scientific misdemeanour, namely that odd practice of attempting to sneak flawed rubbish into the scientific literature in support of dodgy agendas. The curious examples of the misuse of the editorial system that resulted in the retraction of the Wegman paper [ http://retractionwatch.wordpress.com/2011/05/17/controversial-paper-critiquing-climate-change-science-set-to-be-retracted-because-of-plagiarism/ ], the resignation of the Editor of Remote Sensing [http://retractionwatch.wordpress.com/2011/09/02/editor-of-remote-sensing-resigns-over-controversial-climate-paper-co-author-stands-by-it/ ], and the resignation of the editorial board of Climate Research some years ago, are examples. In these cases non-anonymity of reviewers might have quelled these particular abuses.

        chris

        March 23, 2013 at 10:45 am

        • @chris: you have the answers for your questions, right? In my opinion, disclosing names may help back-scratching phenomenon – i presume that this is what happening in the big journals. Do you think good reviewers are agreeing to review papers these days? Editors are finding it difficult to recruit reviewers – especially smaller and open journals.

          Ressci Integrity

          March 23, 2013 at 9:16 pm

  11. Have not read all comments but think it very strongly depends on field.

    I come from the land of protein vibrational spectroscopy. Going into medical diagnostics was a bit of a career risk. But, I wanted to see Europe, and I honestly thought the Germans were on to something. (Never had chance as girl to visit as parents too poor to send me on EuroTrip.)

    Open is fine for math & physics and is already headed that way.

    Chem is heavily invested in patents. This can be good and bad for open review.

    Biomed = huge and needs tons of parallel processing right now. It needs open refereeing — and an educated populace that can understand what it is funding.

    Allison L. Stelling

    March 22, 2013 at 7:54 pm

  12. I think this is a supremely bad idea. In the U.S., a certain quasi-religious organization follows a policy of “lawsuits to harass,” which serves quite well as a disincentive to public criticism of their institution. The present proposal seems likely to be more of an incentive to false witness than an editorial disinfectant, and would likely serve to stifle speculative science and unconventional ideas. The grant review study section system already does this quite well. I’m not sure that adding Klieg lights to the mix serves any constructive purpose.

    The phrase “disclose and speak to the project’s full methods and data” seems especially susceptible to abuse. Perhaps you meant “publication” instead of “project?” Would tangential or incorrect data that led a researcher towards a publishable idea fall under scrutiny as well? How many controls would have to be run for a failed, but inspiring experiment?

    Nigel Tufnel

    March 22, 2013 at 10:29 pm

    • Are you referring to the case of Raj Persaud? Read the wiki page, it’s quite interesting. He was found guilty of plagiarism though it seems he had obtained informal permission from the guy he copied from. More importantly, here is what Wiki says, and what illustrates the danger of these kinds of proposals: “It emerged in the judgement that the matter was brought to the attention of the GMC by the Citizens Commission on Human Rights, an organisation founded by senior members of the Scientology movement.” You can easily see how this same kind of action could be used against anyone who writes something that annoys some organization with deep pockets. Without serious safeguards for this kind of harassment by competitors etc, there is no way schemes such as the one proposed in this thread will get traction.

      Average PI

      March 24, 2013 at 7:29 am

      • Persaud is a great example of the phenomenon that worries me. That organization and others (think gun violence and climate change) have a track record of using “the system” to their advantage, ethics be damned. People only have one reputation, and some questions cause irreparable damage when asked in public, regardless of their foundation in truth. The list of political third rails and economic incentives to quash science is pretty long, and those who would happily abuse a process like that proposed here is even longer.

        Why would one spend money or even risk bad publicity (not to mention jail) for trying to stop research you oppose, when all you have to do is arrange to threaten a few authors with this extortionist process?

        Nobody looks good under the interrogation lamp, even those with nothing to hide.

        Nigel Tufnel

        March 24, 2013 at 5:16 pm

  13. Update on the story mentioned by ferniglab above. 5 cases of duplication of data (from the same lab) documented on my blog. In one of the papers, the entirety of the experimental figures contain data re-use (from 2 different papers). However, correcting the scientific record seems like a long and almost impossible task!

    Raphaël Lévy

    March 23, 2013 at 5:33 am

    • What I’m not doing is proposing an all-purpose solution. What I’m proposing is that the editors of journals, employers, courts etc, are entitled to ask, post-publication, questions of the (principal) authors, and that these questions, and the answers to them can be forensically recorded and archived. A technical advisor might also be on the line.

      The uniform requirements already state (probably to no great advantage) that authors should be able to speak to the data, and this is part of the implied agreement by which papers are published in the great majority of journals.

      If you look at recent cases monitored by Retraction Watch, I think you’ll see several researchers who would not welcome such a process, and might well not cooperate. They might realise what a hostage to fortune it would be to say anything at all. In which case, their paper could be retracted, quickly and, most crucially, cheaply. There would be no grounds for any lawsuit, because no fault is implied. It would be an administrative retraction, purely following rules that say if the author doesn’t speak to the paper, then it will be pulled.

      As it is, when papers come under suspicion over integrity issues, I think very often editors and institutions either duck for cover because they don’t want to get bogged down in some sort of massive inquisition, or they have a “chat” with the authors. Such chats go nowhere. And if the burden is too great, concerns will be buried.

      So I’m proposing something quite simple.

      Brian Deer

      March 23, 2013 at 1:54 pm

      • Brian,

        I am delighted a non-scientist is doing this. You are the perfect choice, and it is good to see you have followed Retraction Watch as I have. It is an exciting time.

        You may note the anonymity of science-fraud whistle-blowers came under considerable attack from many seeking to unveil the individuals exposing science-fraud. Science-fraudster is a case in point. Look at the potential damage the science-fraudsters have tried to do to a truth-seeking scientist who in only 6 months exposed multi-million pound science-fraud around the World. That’s one enlightened individual working for nothing for something he believes in.

        Now we can see why those in favour of removing anonymity of whistle blowers of science-fraud may not be in favour of removing the anonymity of the secretive process of reviewing papers and grants. We need to encourage scientists whom review and reward grants, largely based upon a researchers publication history, to show due diligence. A good starting point would be to remove anonymity. Afterall, they have nothing to hide, right?

        Stewart

        March 23, 2013 at 2:43 pm

      • I suggest to have the same system in place for investigative reporters, especially in the UK (given the recent abuses we all have read about): Government officials and experts should go over all their past pieces and retract those for which the reporter was unable to provide exhaustive evidence for every word in the piece.

        Average PI

        March 23, 2013 at 5:20 pm

        • What happens if the editor (official) is corrupted? Say,for example, editors who ignore, or make light, of complaints about scientific irregularities others can see in the journals they edit.
          Once you have corruption inside the system it is very difficult. That is not a reason not to try to improve though. You can write to COPE, but that does not have any powers.

          http://publicationethics.org/files/u7140/FlowchartComplaints_rev_July172012_ComplaintsOfficer_0.pdf

          fernando pessoa

          March 24, 2013 at 8:42 am

          • There is always the possibility that somebody is corrupt in the chain, that is life. You won’t get rid of that any more than you will be able to get rid of vices such as drinking etc. It is quite possible that the people who scream the loudest are the most corrupt, actually, as that would make an excellent cover. There have been many cases in which the strongest voices against a certain activity turned out to secretly be engaging in that very same activity!

            Average PI

            March 24, 2013 at 10:08 am

        • Interesting to note that my post above, where I propose that the same system being proposed by Deer get put in place for investigative reporters got 7 negatives. If it’s so good for scientists, it’s got to be as good for investigative reporters.

          Average PI

          March 25, 2013 at 3:50 am

      • NPG policies on sharing of materials and data (http://www.nature.com/authors/policies/availability.html):
        “After publication, readers who encounter refusal by the authors to comply with these policies should contact the chief editor of the journal (or the chief biology/chief physical sciences editors in the case of Nature). In cases where editors are unable to resolve a complaint, the journal may refer the matter to the authors’ funding institution and/or publish a formal statement of correction, attached online to the publication, stating that readers have been unable to obtain necessary materials to replicate the findings.”

        That seems not so different from what is proposed here? A formal statement of correction is of course a little short of a retraction but it is a process which is simple, on paper, and does not require a full investigation.

        In practice, in the #stripynps story, authors have refused to share data related to a number of papers (including two Nat Mater articles; http://raphazlab.wordpress.com/2013/03/04/request-for-raw-data-authors-and-editors-responses/), editors have been unable (as of today anyway) to resolve the matter, but no such statements have appeared. This is a case in progress though [Editors were contacted 06/02/2013]

        Raphaël Lévy

        March 25, 2013 at 7:04 am

        • That’s correct. There should be some sequence in the enquiry. First, there is presumption that paper is correct (that is when journal cares that all data necessary to reproduce results are in the paper). Only when results cannot be reproduced, the inquiry starts. Or, when there is obvious flaw in, say, logic of the study. But asking for raw data without a cause seems strange.

          pyshnov

          March 25, 2013 at 9:05 am

  14. If you read the policy on retractions of the publisher, say – Wiley, you will clearly see that the policy is designed to put the decision in the hands of the university/institution. Editors, publishers, all govt. committees/offices on research integrity do the same, and, as a result – universities investigate their own fraud.

    Here, http://authorservices.wiley.com/bauthor/faqs.asp#policy Wiley says:

    “If the Editor of the journal believes there are irregularities with the content of the article then it is at the Editors discretion to print an appropriate statement in the journal. This can only take place after full investigation of the incident has taken place and the nature of the irregularity established beyond all reasonable doubt.”

    Full investigation? But in another place, http://authorservices.wiley.com/bauthor/faqs_copyright.asp Wiley says:

    “Dealing with misconduct
    it is the duty of journal editors to investigate suspected cases of misconduct. They need to decide whether it is necessary to retract a published contribution and in some cases, whether it is necessary to alert the employers of the accused author(s). Some evidence is required, but if the employers have a process for investigating accusations, it is not necessary for the editor to assemble a complete case as this may entail wider consultation which would bring the author into disrepute before the facts of the matter have been decided.”

    It means:university will investigate its own fraud. All other investigations are not independent investigations, they are sham.

    I don’t know how many people realise that the system is rigged so badly. Again, I explained this on http://sciencemishandled.wordpress.com/about/

    pyshnov

    March 23, 2013 at 2:57 pm

    • The reason being, of course, that Wiley doesn’t want to pay for inquiring into its past publications. It wants to offload that burden onto the institution, leaving the journal’s staff to get on and make money for the publisher.

      And, as you say, the institution doesn’t want to investigate itself, much less pick up a million buck bill for the pleasure. University College London spent three years obstructing an investigation of mine, and when the journal finally pressed it to conduct an inquiry into my findings, it said that too much time had passed and it didn;t have authority over the researchers anymore, so nobody would be expected to cooperate.

      Which is why I think progress might come from a simple, administrative procedure. If authors won’t speak to their work, as per the present uniform requirements, then that would be grounds for retraction.

      Brian Deer

      March 24, 2013 at 12:47 pm

  15. In reply to Average PI March 24, 2013 at 10:08 am

    I didn’t deny “There is always the possibility that somebody is corrupt in the chain, that is life.”
    Starting off with that sounds like a deflection and an acceptance.

    The issue is what to do about it.

    Things like corruption can be reduced. You can do something. Countries do change their positions in the transparency index rankings.

    http://www.transparency.org/cpi2012/results

    “There have been many cases in which the strongest voices against a certain activity turned out to secretly be engaging in that very same activity!” Would you like to expand on this? Which activites, which cases for example?

    Even if this were the case it does not take away from the people who trying to reduce corruption.

    fernando pessoa

    March 24, 2013 at 1:13 pm

    • “Even if this were the case it does not take away from the people who trying to reduce corruption.”
      It may not take away from them, but it does not give them automatic moral authority either. They should be scrutinized first.

      Average PI

      March 25, 2013 at 3:48 am

  16. Brian Deer, I want to add something. My view is that every corporation is trying to spend as much money as they can: we don’t live in capitalism, this is socialism because corporations live on public money – they write their expences off the taxes. Everything is moved by politics rather than by the money.

    pyshnov

    March 24, 2013 at 6:09 pm

  17. At this point in science, we have many many thousands of journals. Each journal has many reviewers for each article – between 2 and 3. So, now we will have another layer of review? Who will do these reviews?

    StatObserver

    March 25, 2013 at 12:00 pm

    • Who will do this review? Good question. And what/who will ensure that those doing the review are not corrupt or have hidden agendas?

      Average PI

      March 25, 2013 at 12:35 pm


We welcome comments. Please read our comments policy at http://retractionwatch.wordpress.com/the-retraction-watch-faq/ and leave your comment below.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 31,183 other followers

%d bloggers like this: