What should an ideal retraction notice look like? We (and COPE) want your input

copeLast week, we announced a new partnership with PRE (Peer Review Evaluation) “to improve access to information about retraction policies.” The first step, we and PRE said, was that Retraction Watch would create guidelines for retraction notices, to which PRE’s flagship product, PRE-val, would link.

Well, it turns out that great minds think alike, or along similar lines, anyway. Today we learned that next week, the Committee on Publication Ethics (COPE) will be discussing a standard retraction form proposed by friend of Retraction Watch Hervé Maisonneuve, who has published several papers on retractions.  According to a writeup:

This template would meet the pre-requisites for the COPE  retraction guidelines[3] using very simple tick boxes: who is retracting the article, the reason for the retraction and history of errata/ expressions of concern. Then, a free text box would allow the editor to add any information they consider useful.

Here’s the proposed form.

While our guidelines will probably describe what should go into a narrative, instead of form a checklist, and we have a few more items that we think belong in a notice, we can certainly appreciate the appeal of a form that reminds editors what may be missing as they’re working.

COPE is seeking feedback on the proposal, which you can leave here or as a comment on this post, since we’ll be adding a link as feedback.

And we’d like to hear suggestions for our in-process guidelines, too. One commenter left some smart ideas yesterday, so please keep them coming in the comments.

14 thoughts on “What should an ideal retraction notice look like? We (and COPE) want your input”

    1. I have some queries:
      a) how will this new “standard” be applied by all publishers in a standard way, or will this be specifically for COPE “members”? Assuming that the RW-PRE version is different to that of COPE, how will the two be integrated and implemented?
      b) what rights do authors have to refute claims, who exactly writes the wording of the retraction notices and who verifies the final content?
      c) what time frame is give to authors, editors and the publisher to complete information before a retraction notice is issued?
      d) Will a retraction notice be forcefully published if not all authors agree with it? In other words, will the publisher trample upon authors’ rights to disagree?
      e) Will this new “standard” be applied only from the day it is launched onwards, or will it also be employed retrospectively to correct all retraction notices that have dsmally failed thus far to explain the background behind the retraction?

      As an example of e), I give two retraction notices in two Springer Science + Business Media journals (note Springer is a COPE member), whose retraction notices are abismal in transparency and states only the following (the exact same wording for both notices):

      “The article is retracted due to disputed authorship and severe omissions in the information of collaborators about the publication.”

      Retraction 1:
      Acta Physiologiae Plantarum July 2010, Volume 32, Issue 4, p 821
      Date: 18 Feb 2010
      RETRACTED ARTICLE: ASYMMETRIC LEAVES2-LIKE16 gene, a member of AS2/LOB family, is required for lateral root formation of Arabidopsis
      Guang-Chao Wang, Lai-Sheng Meng, Xu-Dong Sun, Zhen-Hua Feng, Jian Zhu

      Retraction 2:
      Plant Cell, Tissue and Organ Culture (PCTOC)
      August 2010, Volume 102, Issue 2, p 265
      Date: 26 Mar 2010
      Retracted Article: Overexpression of the Arabidopsis ASYMMERTIC LEAVES2-LIKE11 gene in Cockscomb leads to modification of flowering and lateral branching
      Xu-Dong Sun, Lai-Sheng Meng, Guang-Chao Wang, Ying Zhou, Zhen-hua Feng

      Why can we not see the open access paper of the retracted paper as COPE requires (with a red stamp on it), even though Sringer is a COPE member?

      And what is the link between Retracted paper 2 and this paper?
      Plant Cell, Tissue and Organ Culture (PCTOC)
      May 2010, Volume 101, Issue 2, pp 193-200
      Date: 14 Feb 2010
      ASYMMETRIC LEAVES2-LIKE11 gene, a member of the AS2/LOB family of Arabidopsis, causes pleiotropic alteration in transgenic cockscomb (Celosia cristata)
      Xu-Dong Sun, Lai-Sheng Meng, Zhen-Hua Feng, Jian Zhu

      Why should one have to purchase the PDF file of these retracted papers for $39.95 / €34.95 / £29.95 when COPE indicates that the retraction notice should be open access?
      Case 1:
      Plant Cell, Tissue and Organ Culture (PCTOC)
      April 2012, Volume 109, Issue 1, pp 91-99
      Date: 04 Nov 2011
      RETRACTED ARTICLE: The AtCCX1 transporter mediates salinity tolerance in both Arabidopsis and yeast
      Zhiyuan Chen, Yongjun Wu, Lijun Di, Guodong Wang, Yufang Shen

      Case 2:
      Plant Cell, Tissue and Organ Culture (PCTOC)
      April 2012, Volume 109, Issue 1, pp 91-99
      Date: 04 Nov 2011
      RETRACTED ARTICLE: The AtCCX1 transporter mediates salinity tolerance in both Arabidopsis and yeast
      Zhiyuan Chen, Yongjun Wu, Lijun Di, Guodong Wang, Yufang Shen
      (although there is a separate site on SpringerLink if one searches hard enough, that lists the open access version of the retraction note:
      http://download.springer.com/static/pdf/634/art%253A10.1007%252Fs11240-012-0193-y.pdf?auth66=1411057037_795d3319eea4899425d50ea7ef6cd880&ext=.pdf (and when we read that notice, we see that the information is so poorly explained that it does not even indicate what the “formerly published article” actually is).

      Case 3:
      Why is the reason for retraction listed as a potential misconduct, and why is the language of the retraction notice so poor? “This article has been retracted due to potential misconducts mainly concerning manipulation and repeated uses of hotomicrographs of control data internally along with mislabeling and/or externally in multiple publications.”
      Molecules and Cells
      July 2012, Volume 34, Issue 1, pp 43-52
      Date: 13 Jun 2012
      RETRACTED ARTICLE: Nicotiana tabacum Tsip1-interacting ferredoxin 1 affects biotic and abiotic stress resistance
      Sung Un Huh, In-Ju Lee, Byung-Kook Ham, Kyung-Hee Paek

      So, may I suggest to COPE to please first request its paying members to please enforce these rules first before it tries to enforce it on the globe’s scientists. Secondly, please go back and make sure that all your member journals and publishers fix all the past retraction notices, because they are totally abismal in most cases.

  1. 1) It’s great to have all data condensed over a single page.

    2) In the pdf available via the RW link, some Tab stops in the “Retraction type” section are wrong, making the end of this section a bit confusing. For instance, the “Authorship” box appears in the middle between two columns, while this box is obviously expected in the “Not invalidating the results” column. Maybe this results from a wrong Adobe configuration for Letter/A4 paper sizes.

    3) Regarding the “Invalidating the results” column, I’m concerned that fuzzy frontiers between “Fraud”, “inconsistent data” and “honest error” might generate insurmountable difficulties of interpretation. For example, inconsistent data may be introduced in a paper intentionally (fraud) or unintentionally, because of a transcription problem (honest error). On the other hand, consistent data coupled to a wrong interpretation may also be considered as fraudulent or not. Will such issues be addressed before completing the form?

    4) If nobody has to sign the form, how to know who completed it?

    5) Is the form intended for both “retractions” and “partial retractions”?

    6) Possible typo (I may be wrong): maybe “Retraction Date” should read “Retraction date”.

    1. The form is a bit weird. Who will fill it out? Who will decide what to tick? Who will ever admit fraud in the first place? Best to put the list of issues to authors and have them make point-by-point statements on how these happened. Let the readers decide how credible the explanations are. And ban certain expressions, like “despite all most of our data being manipulated or invented, our main findings remain valid”.

      1. If the authors refuse to answer to the list of problematic issues, so be it. The journal should publish the list and the statement: “The authors refused to comment” as part of retraction notice. Again, let the readers decide why the authors may have chosen not to comment. But all issues, which made the paper untenable, must be listed by the editors.

  2. I would like to emphasize that with the retraction of any plagiarized paper there should be not only a major, official excuse to any of the potential victims and the payment of a penalty, for example 100.000 USD for every year of not admitting the plagiarism.

    1. Eibl, enforced by whom? So, how would transnational cases work, or are you suggesting that extradition is the solution if the fine is not paid? Where did you get this sum of 100K US$ from? I think what you are discussing is irrelevant to the wording that the retraction notice requires. Instead, you are dealing with a downstream process.

    2. If you mean that the fine should be directed to the plagiarized author(s), please feel free to plagiarize all my papers, without admitting anything for many many years.

    3. Major professors should be punished by the scientific community. Any plagiarism supporting institution should pay to the victims of plagiarism – it appears to be logic that some universities try very hard to not to stick to their own rules, for example, if a preliminary investigation within 14 days is guaranteed, this should not take about three years to get over any crimes by a university.

  3. This is what should NOT be part of a retraction notice: “Our key findings remain valid”. Also, JBC’s notices consisting of “The article was retracted by authors” and nothing else should also not be acceptable. Otherwise, it would be nice if authors would honestly state what exactly makes their paper untenable.

  4. A good example of an ideal retraction notice:
    “I hereby retract my paper due to a fatal error I explained in [2]. All my attempts to patch the error have failed. I thank Shi-Lei Su, a student from Yan Bian University, Ji Lin Province, China for bringing the error to my attention.” (Mladen Pavičić, see http://retractionwatch.com/2012/11/30/poignancy-in-physics-retraction-for-fatal-error-that-couldnt-be-patched/#more-10910
    A very clear description from the author himself in which he describes why this paper needs to be retracted. On top of that with a clear acknowledgement to Shi-Lei Su. Shi-Lei Su is able to add this to his CV.
    http://www.sciencemag.org/content/345/6202/1254.1 is another [recent] example of a very clear retraction note.
    Prepared by all authors and with a very extensive description why the paper needs to be retracted See also http://retractionwatch.com/2014/09/15/hayabusa-science-retraction-made-official-but-behind-a-paywall/

  5. The advantage to a checkbox-based approach is that it would be machine-readable. If the function of a reaction is (a) to inform the audience and (b) to prevent the paper from being cited, objective (b) is more easily achieved if literature storage/search systems can easily process retraction notices and use the information to exclude retracted papers or by default display retractions rather than the papers.

  6. According to this article, the statement “Today we learned that next week, the Committee on Publication Ethics (COPE) will be discussing a standard retraction form proposed by friend of Retraction Watch Hervé Maisonneuve” was made in September 2014.

    The COPE Guidelines for Retractions PDF file is still dated September 2009:
    Elizabeth Wager, Virginia Barbour, Steven Yentis, Sabine Kleinert

    When can we expect to see a new set of guidelines by PRE and/or COPE? Or, if such new guidelines already exist, can someone kindly post the URL. Thank you.

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.