Poll: Should co-authors of a scientific paper be allowed to change their minds?

lancetSince we reported Friday that multiple authors had asked to remove their names from a high-profile 2011 Lancet paper about a risky transplant surgery, a few readers have wondered: Should this be allowed?

To recap: The same day the journal announced it was tagging the controversial paper with an expression of concern, it issued a new erratum about the paper, removing three author names (one had already asked to be removed earlier). The highly cited paper has been under scrutiny ever since the last author, Paolo Macchiarini, has been facing allegations of misconduct, which most recently led to Macchiarini’s dismissal from the Karolinska Institutet. (Here’s our timeline of events to keep you abreast.)

It’s not surprising that a few of Macchiarini’s co-authors would want to distance themselves from this ever-expanding scandal, but should authors who originally signed onto a paper be able to change their minds? Let us know in our poll, below.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

36 thoughts on “Poll: Should co-authors of a scientific paper be allowed to change their minds?”

  1. I voted “Only in specific circumstances”, but I think the circumstances should be limited.

    Basically, I think an author should only be allowed to withdraw authorship if new information has come to light, which

    a) Had they known it, they would not have consented to be an author, and
    b) They didn’t know at the time (before agreeing to be an author), and
    c) They could not reasonably have been expected to know at the time.

    Point “c)” is key, I feel. Authors have a responsibility to take reasonable steps to verify the soundness of the science they put their names to.

  2. I have been banging on about this for ages in my blog: To many so-called authors are only too happy to be in for the glory but deny all responibility, when things go belly up. You either are an author or you aren’t. And if you brashly claim to have written the thing, should you not at least have read it before it’s published? Those over long author lists are ridiculous anyway. If all you did was supply a specimen from your store, your rightful place is with the thanks and acknowledgements – it’s honest and it’s safe.

  3. Axel, I think there is a word for this: accountability. Authorship, and all its glory, also implies accountability. Trying to remove one’s name implies trying to avoid being accountable. Most journals, upon submission, require that the corresponding author make a statement something to the effect of “all authors have agreed to this submission”. However, Neuroskeptic may be right, in this Macchiarini case, that c) may apply. Even so, a guarantee at the time of submission was clearly false and/or violated. In such a circumstance, surely the paper should be retracted? I do not have a copy of the said paper, but what does the authorship contribution statement state? If it indicates that some of those who wish to retract their names, but they have been involved in experimental design or execution, then removing names is not compatible with a simple erratum.

  4. I voted “never”. That said, I would allow it in some very rare circumstances, for example if someone was included in the list of authors without his/her permission. But I don’t think misconduct by a co-author should allow removing one’s name.

    1. Anon, I voted and commented precisely in the same way. From another angle, this is how I see it: when the paper was accepted for publication in The Lancet, the team (i.e., 100% of all authors) rose and benefited (after all, it’s not every day one gets a paper published in The Lancet). So, when one falls, they fall collectively as a team. Those who voted yes seem to be confusing science publishing with soccer. Here’s the difference: when one get’s a red card, unlike soccer where only one player gets sent off, in publishing, the whole team gets kicked out. Let’s not start changing the publishing landscape into some soft form of democratized soccer game.

  5. I voted “never”, because if you want the credit of a paper, you should be responsible for what you have published.

  6. I disagree with most commenters here. I think removing yourself as co-author is the best option if you would like to retract the paper, but the corresponding author refuses to do so. I gives an even stronger message than the expression of concern that the paper is fundamentally flawed and all that stands in the way of a retraction is the stubborness of the lead author(s).

    1. Geologist,
      although it looks like contradicting myself, I agree with you. That is, if as in your example the co-author notices the problems first and does it to blow the whistle. That’s not what normally happens. Everone is happy until the thing goes belly-up. Only then do the rats try to disengage from the sinking ship, and that’s the time, when, IMHO, they should no longer be allowed to do so.

  7. I voted ‘allow in specific circumstances’ and in general agree with Neuroskeptic. To all commenters taking the hard line (‘you approved it, you are fully accountable’) I can only say that this is simply impossible in the era of multidisciplinary science.

    To start with, I don’t think it is fair to expect anyone to be an expert in everything – if we could be, we would still live in the age of one-author papers. There is a reason why we don’t. Papers today can be very complex, include highly sophisticated computational analyses and very involved wet-lab experimentation. I think it is probably reasonable not to expect from the author, whose contribution to the paper is, for instance, generation of biological model, to be able to perform or even understand (beyond basic) some more difficult computational analyses.

    And so, if the discovery of the issue with data fabricated by one dishonest authors requires e.g. near forensic data science expertise, is it fair to ask the other authors to bear the brunt? Unfortunately, it is very easy to take a moral high ground if it never happened to you. But of course I agree that there should be limitations to when we would allow the authors to distance themselves from the study, and I suspect such situations are probably best judged on case-by-case basis.

    Two more points:
    First, the discussion, inevitably and notably, comes back to a common misconception of what is the purpose of correction/retraction – and it is not to penalize the authors, but to ensure that the scientific record is correct.

    Second, given the voices asking for public humiliation of all authors, because when they published in Lancet, they benefited from it, I wonder if you would be as full of condemnation if we were talking about correction/retraction of a paper in a low impact journal? Just curious.

    Competing interest: I am an employee of BioMed Central. The opinions expressed here are my own.

  8. I agree with Misha, if you took credit of the paper then how could you shed off the responsibility when it is retracted.
    While accepting authorship one has to be fairly sure of what has gone in the paper and take onus of it. If things have gone wrong it is an oversight by all authors (and not just corresponding author). I understand that in cases of multi-disciplinary papers it is difficult to judge the authenticity, but most journals allow to cite where the data has come from and which author is responsible for what work. I guess the problems arises because we dont take the authorship guidelines seriously and in my humble opinion the “gift” authors are perhaps ones who want to shed off the responsibility.
    I am sure the coauthors of the controversial paper must have based in its glory for years and taken benefits of it.. and now in times of distress they are leaving him alone.. its simply disgusting to see it.

    1. But the situations we are considering for allowing this are different – we (Neuroskeptic, me) refer to cases when one of the co-authors was genuinely dishonest. Data – any data – can be fabricated, and if you really wanted to, you could make it indistinguishable (or barely distinguishable) from real data. I think it is a fair assumption that researchers trust their collaborators enough not to go over every piece of data with a magnifying glass. Should they be blamed for being to trusting? Perhaps. Should they be punished for it? I cannot see why – because I cannot see why they should be punished for one dishonest person being a good forger.

      P.S. And given the language we are using here, we also go again and again back to the issue of what is the purpose of corrections/retractions…

      Competing interest: I am an employee of BioMed Central. The opinions expressed here are my own.

      1. But in such a situation, shouldn’t the paper just be retracted? The journal does not need to have the consent of all coauthors to retract the paper, so if very serious doubts surface, why wait for consensus among the authors? But if you have hard evidence that something is seriously wrong you can also write a comment on the paper (although not all journals accept those of course). Expressions of concern. Publisher permitting, there is artillery available to attack papers, you can use it against your own papers as well, without distorting the original in the process.

        Besides, if authors want to to distance themselves from a paper, should not an absolute minimum requirement be a clearly written statement detailing why the authors no longer wish to be considered part of the paper? A proper “retraction-of-my-name” notice by RW’s guidelines for retraction notices is the very least I would ask for. The reason we got from the Lancet was that the authors “no longer wish to be associated with this paper and ask for their names to be removed”. Huh!

        Concerning the purpose of retractions: The reason to retract is of course to correct the record. Myself, I think this shame of retraction is a thing that will pass. The more retractions there are (and they seem to be constantly going up), the more RetractionWatch and other report on them (which they do) and the more we discuss retractions (which we do, at length), the quicker we will move to a system where the retraction is considered only a somewhat more embarrasing version of an erratum and misconduct is kept track of in some separate system.

        1. I quite agree with you on all counts! To your question of whether the paper should not just be retracted, it is so wrong that authors do not feel the conclusions are supported once you remove fraudulent data, I think the answer is what somebody mentioned before – that other authors and more importantly the editors may be the ones “blocking” retraction (or even just a correction). Retracting your name is something that may be easier for many to swallow – but I really like your idea of Retraction-of-name.

          And yes, misconduct should be controlled otherwise – and it is, that is what research integrity institutions are for.

          Competing interest: I am an employee of BioMed Central. The opinions expressed here are my own.

          1. You are probably right, removal of names feels like a “lightweight version” of a retraction and this is much of its attraction on all sides. I just don’t agree that it is a very lightweight action, authorship is normally considered very important (authorship disputes seem to be a rather common cause of retractions, for instance).

            If the editorial office want to flag that misconduct is suspected they should use some more transparent method than to allow people to remove their names. If they do not want to flag it, then there is also no reason to remove people from the paper. If misconduct is confirmed there is no reason to do anything but retract it. The problem that editorial offices are reluctant to discuss problems with papers and too slow to retract is not solved by giving them an easy and even less-transparent back door to use while they think about it.

          2. Torbjörn
            > If they do not want to flag it, then there is also no reason to remove people from the paper.
            I think you’re missing part of the point here. You detect a falsification, but the other authors insist all is fine. A check by the editors will take time and they may or may not retract – out of your control. So what to do in the meantime? You can distance yourself as soon as you see reason to and before anyone else notices anything. In that case whatever happens you have done the decent thing and you’re – rightly – in the clear. If you keep shtum and wait if perhaps nobody else notices anything, then you rightly sink with the rest.

          3. Axel
            > A check by the editors will take time and they may or may not retract – out of your control. So what to do in the meantime?

            I see a difference in our viewpoints here… I see removal of names (at request) mostly as a way to temporarily release som pressures on the journal, not a service to the authors. Everything that happens to the paper post-publication is out of control for the authors, the journal is completely in charge no matter what. If we decide that journals should allow this in exceptional circumstances, we’d just subdivide the retraction question into smaller pieces. I saw this as a question about what the journal should ideally do – and I think that is not to remove names, it is to work as quickly and transparently as possible to a clear decision of whether to retract the paper or not.

            Meanwhile, the authors have a whole Internet full of communication channels if they want to distance themselves from a paper. Myself, I would just write up the story, including all my evicence, and publish under my own name on PubPeer (questions of anonymity would obviously no longer apply at a point where revoking my coauthorship is an option…). That would put my (timestamped) seal of disapproval on the paper for the while being.

      2. I agree with you that we trust our collaborators and in all good faith the data is shared and its almost impossible to find out fabrication well until the cat out of the bag.
        Now the point is if collaborators X and Y gave genuine data, Z who is a main author had fabricated. I fail to understand why X and Y should distance from Z. If I were X or Y I will say Figs A B and C has data from my lab and I stand by it. But since Z has fabricated and if my data does not make sense in absence of Zs data, I will recommend a retraction, else I can not and should not. Let the editors decide. I dont think journals should allow authors to abandon the ship as and when they feel like.

        1. Deepak
          > I will say Figs A B and C has data from my lab and I stand by it.
          This is not about one or two isolated figures, it is about an article in its entirety. If you happen to come across falsification, you must clearly say so at once and distance yourself from authorship. Retraction or not is for others to decide.
          What you describe is valid for another article in a collection you contributed to, but not an article under your co-authorship.

  9. I voted yes, since there could be cases, neither a senior author, nor the editor want to retract a paper, and many of the co-authors did not know about scientific misconduct, like plagiarism, or not acknowledging the originator of a project, who invested several years of unpaid, highly independent and extremely pioniering work – and then see himself not as co-author of his own project.

  10. As one of those who cried for blood in an earlier comment, I should probably comment on my choice here… as well as expand a little on the not-too-well articulated thoughts I had earlier.

    I voted Never, and the Imposing Factor of the journal has absolutely nothing to do with it. But I think the situation with the current Lancet paper nicely illustrates why this should be Never, and offers a suggestion as to what should be done instead: If the request for author removal is reasonable, then retract the paper and then reconsider it after suitable revisions (such as removal of any section or result which no longer has a contributor).

    With the Lancet paper, it is no longer clear whether the all content there has anyone responsible attached (apart from the overall general responsibility of everyone and the heavier responsibility with “lead” authors). Just as Rafal says, it is often impossible for everyone on a paper to have full grasp of all aspects of the work, this is precisely the reason why we cannot have authors suddenly abandoning ship, they may have contributed key data that the others can’t fully answer for.

    Papers are reviewed and eventually accepted based on among other things the stamp of approval that suitable expertise attached to the project will lend. The Lancet has now removed four names and the corresponding author contributions. Would the paper have been accepted in the first place without these names and their contributions? I don’t know, I can’t even see what their contributions where anymore (in this particular case I would be clueless anyway since I’m a physicist with no domain expertise whatsoever, but that’s beside the point). I can’t see any other sensible way to handle this problem generally than to retract and redo the whole submission from scratch to find out.

    This is not really a very drastic suggestion. Just as geologist implies above, with an attached statement of concern and four stepped-down authors, the paper is retracted in all but name. So why not just take the last step?

    1. Ha, thanks for this explanation. I will admit you are making much more compelling case for Never than “there must be blood”! I am even considering revising my position (but does RW allows corrections? ;))

    2. I think you make a very good proposition. Too many things are hyped when published and then thrown and disowned without proper assessment of what holds and what does not. For the co-authors them-selves it seems better to take that approach, it is also in their interest that of the scientific community.
      But generally speaking the reporting of science still needs to be changed, too much depends on big ego, internal fight, financial interest, in the end who cares about understanding, improving, curing?

  11. From the paper.

    “Contributors
    PJ was responsible for the bioreactor-based cell seeding; assisted the surgery and with collection of secondary data; and wrote corresponding methods, results, and interpretation sections. EA and TS undertook all flow cytometry characterisation of the cells, interpreted the results, and wrote corresponding methods. SB provided preclinical data for human tracheal biomechanics, and helped to write the report. KLB, BN, and GM designed, undertook, and assessed the multiplex analyses and wrote corresponding methods. KLB undertook the bone-marrow isolation. PB organised and supervised standards at the good manufacturing practice facility. BB did histological evaluation. CC and AS designed and developed the three-dimensional nanocomposite trachea, and wrote corresponding methods. OE and TG are responsible for the clinical follow-up of the patient and provided biopsy material and blood samples; TG also participated in the surgery and wrote corresponding methods. K-HG and JL assisted the surgery. SLG and SS did all cell imaging and wrote corresponding methods. OH and TL undertook and assessed all epigenetic analyses and wrote corresponding methods. JEJ and GH assisted in the preoperative and postoperative care. BL did all radiological imaging, interpretation, and three-dimensional reconstruction, and wrote corresponding methods. TL and CR undertook and assessed the miRNA studies, and wrote corresponding methods. VL and AIT undertook and analysed gene expression experiments, and wrote corresponding
    methods. EW isolated the mononuclear cells. PM was the primary investigator and leading author of the report, indicated how to built the three-dimensional nanocomposite, was leading surgeon and was responsible for the preoperative and postoperative course, and oversaw the review process. All authors provided primary data for modelling scenarios, and assisted with interpretation of results and report revision.”

    Do all authors fit ICMJE requirements for authorship?

    1. Key snippet in this contributorship statement that may have been overlooked by some: “PM…oversaw the review process”

      Does this imply that only PM engaged with the peers and editors, and that all authors were not aware of responses to peers/editors? It would be interesting to learn the opinion from all authors about not only the research aspects and the paper compilation, but also about the participation of each in the peer review process.

      In addition, I wonder if this paper had been submitted elsewhere, and rejected, before it landed up in The Lancet? Finally, given the massive public concerns and media attention, would it not be wise for Horton to release the peer review reports, to make sure that the whole process is fully transparent?

    2. Just to tag things properly: This is the original Contributors section, without names removed. Thanks for posting!

      As for its interpretation, I suggest to look at the external evaluation, downloadable from somewhere at RW (and from KI). I think that Gerdin had access to that communication.

      Another fact here, to highlight the complications with removing author names: the first author to get his name removed (K-HG above) was also heavily criticized in the external report, based among other things on these descriptions of his contributions in the questioned papers.

  12. Rafal
    I voted ‘allow in specific circumstances’ and in general agree with Neuroskeptic. To all commenters taking the hard line (‘you approved it, you are fully accountable’) I can only say that this is simply impossible in the era of multidisciplinary science.

    I think it should go as follows:

    Basic principle: retract a paper that is troubled; demand a concrete retraction note that clears anyone to be cleared.

    If the journal dclines to retract; then demand your name to be removed.

  13. I agree with Neurosceptik. Remember the Wakefield paper, in the end only Lancet and Wakefield supported the paper that turned out to be a fraud. Everybody new it already but it took Lancet 12 years to retract that paper. A lot of patients were harmed by that.

  14. I’d like to know who established/decided that a retraction should not be about punishment. When did this rule become implemented? Is this a COPE rule or a generally accepted rule?

  15. I voted No
    It is true that in some papers may be impossible for authors to understand properly parts produced by other authors. Big physics papers (Higgs’ boson, LIGO) probably fall into this category. However, all authors should be able to stand up and give a talk about the paper, not just focussing on their area of expertise, but also presenting the Big Picture. So as authors we are are all in this together. Now, should it transpire after publication that part of the paper are not founded on real data, then all that needs to happen is that one’s own part is re-interpretated.

    The new interpretation could be published elsewhere as a solo paper, especially nowadays, when there are numerous open access and open data platforms for this. The new interpretation and associated data provide the community with evidence to understand what may be the problems with the original paper, which may in due course be retracted. I think this is a much better way forward than jumping off a paper. Jumping off a paper simply reflects a motivation driven by a wish for glamour, Imposing Factors and so on.

  16. Perhaps it goes without saying, but another circumstance in which an author should be able to have their name removed after publication is if they never agreed to be a co-author in the first place.

  17. So how does one cite the authorship of a paper that evolves over time? Disqualification of one team member implies automatically disqualification of the whole paper.

  18. I vote No.
    In this specific case why close collaborators (PhD student for instance) who were under his/her responsibility are left in the author list!
    All authors have be responsible according to their involvement of course, for the best and the worst.

  19. Anna Hörnlund, Chief Legal Counsel at the Swedish Research Council, published a stetement on April 30, 2016, in the Lancet:

    “On Jan 1, 2004, a law came into force in Sweden concerning the ethical review of research conducted in human beings. This law covers research conducted in living human beings, on human cadavers, and on biological material from human beings, and the handling of sensitive personal information. The Swedish Research Council considers Paolo Macchiarini’s activities1 to be research conducted in human beings.

    When research is conducted in human beings, the principal investigator (defined as the state agency or the physical or legal entity under whose organisation the research will be conducted) is obligated by Swedish law to apply for an ethical review. The application must be submitted to one of six regional ethical review boards. These review boards are individual public authorities. Neither Macchiarini, nor the Karolinska Institutet, has submitted such an application.”

    http://www.thelancet.com/journals/lancet/article/PIIS0140-6736%2816%2930320-8/fulltext

  20. Collaborations between mathematicians and biologist will sometimes involve
    1) Biologists who cannot check the mathematics and
    2) Mathematicians who cannot check the biology.
    High quality research will sometimes require interdisciplinary collaborations, and I cannot see how scientists can be held responsible for contributions they had no part in. However, if a scientists discovers that his/her publication depends on fraud, then it seems to me that the scientists should be allowed to distance himself/herself from the work. Perhaps that could help revealing fraudulent work faster. The most important thing is to quickly identify false result and make the criticism known, so as to protect patients (or more frequently resources) from the possible damage of false results. A lot could be done for that goal.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.