Retraction Watch

Tracking retractions as a window into the scientific process

Another win for transparency: JBC takes a step forward, adding details to some retraction notices

with 16 comments

jbc 3115Retraction Watch readers may recall that we have been frequent critics of the Journal of Biological Chemistry (JBC) — published by the American Society for Biochemistry & Molecular Biology (ASBMB) — for their opaque retraction notices. Such notices often read simply “This article has been withdrawn by the authors.”

But we are — despite what some might say is evidence to the contrary — eternal optimists, so when the ASBMB announced they were hiring a manager of publication ethics late last year, we cheered. (Patricia Valdez, a former NIH staff scientist, has since filled that position.) And today, we have another reason to say “Hurrah!”: JBC retraction notices will now include “additional details provided by official [Office of Research Integrity] ORI or institutional reports,” the journal tells us.

Here, for example, are five retractions in the March 1, 2013, issue by former University of Kentucky scientist Eric J. Smart, whom the ORI found to have faked dozens of images:

This article has been retracted by the publisher.

An investigation by the Office of Research Integrity determined that falsified and/or fabricated Western blots were included in Figs. 1A and 5B (https://federalregister.gov/a/2012-28209).

This article has been retracted by the publisher.

An investigation by the Office of Research Integrity determined that falsified and/or fabricated Western blots were included in Figs. 2 (A–C) and 4 (A and B) (https://federalregister.gov/a/2012-28209).

This article has been retracted by the publisher.

An investigation by the Office of Research Integrity determined that falsified and/or fabricated Western blots were included in Fig. 2A (https://federalregister.gov/a/2012-28209).

This article has been retracted by the publisher.

An investigation by the Office of Research Integrity determined that falsified and/or fabricated Western blots were included in Figs. 5 and 7 (https://federalregister.gov/a/2012-28209).

This article has been retracted by the publisher.

An investigation by the Office of Research Integrity determined that falsified and/or fabricated Western blots were included in Figs. 1A, 2, A and B, and 4A (https://federalregister.gov/a/2012-28209).

Smart has had one other retraction so far, in PNAS.

Three retractions for Terry Elton, a serial image manipulator, according to the ORI, have also run in the JBC recently. Two include the reasons why:

This article has been retracted by the publisher.

An investigation by the Office of Research Integrity determined that falsified and/or fabricated Western blots were included in Fig. 6 (https://federalregister.gov/a/2012-30866).

This article has been retracted by the publisher.

An investigation by the Office of Research Integrity determined that falsified and/or fabricated images were included in Figs. 2 (C, D, and F), 3 (C and E), 4G, and 5 (C and F) (https://federalregister.gov/a/2012-30866).

One, however, doesn’t. The notice for “MicroRNA-155 regulates human angiotensin II type 1 receptor expression in fibroblasts,” a paper originally published in 2006 and cited 122 times, reads only:

This article has been retracted by the publisher.

What seems to explain the difference is that the ORI notice only referred to the first two Elton papers (and the microRNA paper wasn’t federally funded, so falls outside of ORI’s jurisdiction). We of course would like to see more information in all of the JBC’s notices — but we like to applaud moves toward transparency whenever we can.

Comments
  • G Knight March 12, 2013 at 10:55 am

    To me, it is important to know if the article was retracted for falsification of evidence, plagiarism or simply published twice.

    Obviously falsifying evidence is a much more offense, even criminal fraud where it was used to get a grant. Most plagiarism is not lengthy enough to be criminal. Publishing twice would act to confirm your own results, but anyone seriously interested would look at the author list and see it was not confirmation.

    • G Knight March 12, 2013 at 11:01 am

      Just to be clear, plagiarism of data and lengthy plagiarism of words could well be criminal fraud too.

      Using the same wording as someone else to describe the same phenomenon, even if it was 4 or 5 sentences in a row, nobody outside of academia would take that as a serious offense.

      Whether criminal charges are laid or a guilty verdict rendered, the crime occurs and the criminal is made when the state’s or the nation’s *criminal code* if violated. The individual transforms themselves into a criminal by their own conscious act violating a criminal code. A verdict in court only adds the adjective “convicted”.

      • Sylvain Bernès March 12, 2013 at 1:46 pm

        And in order to make clear your clarification, you should differentiate plagiarism and self-plagiarism. The thief burglarizing its home is generally seen as a complete idiot. However, in some instances, self-plagiarism may also be considered as illegal, because of copyright issues.

        • Marco March 12, 2013 at 4:12 pm

          Well, you could see self-plagiarism as a burglar taking stuff from his own home…and then getting money back from the insurance.

          Self-plagiarism artificially inflates your productivity, and productivity is used to determine who “deserves” a promotion or funding. Not too long ago one of my projects was not funded, even though it was considered fundable and had good reviewer comments, because my scientific production was considered too low compared to those who did get funding. It crossed my mind, just briefly, what a bit of self-plagiarism could have done with that decision…

          • AllOutWar March 12, 2013 at 5:56 pm

            Marco, what if no financial, position, grant or any other tangible benefits are gained? Would self-plagiarism then be allowed?

          • Marco March 13, 2013 at 2:35 am

            AllOutWar, allow me to first explain my view of what self-plagiarism is: it is presenting something as new which isn’t. If you tell the readers that you have published something before and here present it again (for some reason or the other), it is not a problem, and I would not consider it self-plagiarism. If you do not tell your audience, it is in my view inappropriate.

            A problem I forgot to mention is also that many good journals have limited space. They may reject my paper, limiting my ability to get exposure of my ideas, because they accepted a paper that self-plagiarizes an earlier paper (i.e., it is presented as a novel idea, but isn’t). That is again unfair competition.

            Furthermore, an obvious problem is that it will be difficult to prove that there are no tangible benefits. How do you prove that your self-plagiarism didn’t help with getting a grant? Or a position?

  • elledr1ver March 12, 2013 at 12:10 pm

    I still find the JBC website fairly opaque in regard to retractions. If you click on a link to one of the above articles (not the retraction but the actual article), you can barely tell the article has been retracted. There is a small highlighted area on the right that says “A correction has been published”. If you decide to save trees and read the article online (full text but not PDF), you might never know that the article you’re reading has been retracted. Other Journals make the retraction much more apparent on their website.

    • Mr. chromatin March 12, 2013 at 3:05 pm

      When is JBC/ORI going to retract the following two papers,

      Sankaran et al., February 17, 2012, Vol.287, No.8,

      Marzook et al., February 17, 2012 Vol 287, No.8,

      Fig no. 7C from Marzook matches with Fig. no. 8b and 8d of Sankaran.

      Fig no. 7A from Marzook matches with Fig. no. 8c of Sankaran.

      http://www.pnas.org/content/110/10/4147.extract.html

      • Junk Science March 13, 2013 at 1:11 pm

        I hope soon, this is from the lab that had 30-40 papers (and at least five JBC papers) questioned by the SF blog, just before it was put down. PNAS issued only a correction, which is very sad indeed. ORI is hopefully investigating it (or as usually is the case: the authors correct one by one without letting the editors know about all the other papers that have been questioned.)

      • Junk Science March 13, 2013 at 1:58 pm

        And I forgot to mention that the last author of Mr. chromatin questioned papers (and other papers by the same authors in JBC) is a JBC editorial board member, so this will be a test if JBC indeed has taken a step forward.

    • Sylvain Bernès March 12, 2013 at 3:48 pm

      An improvement would be to tag the html version (both abstract and full paper) with a large “RETRACTED” patch, as done in the pdf version. The point is however a minor issue compared to more significant problems related to the persistence of retracted articles in databases. For example, I suppose that Thomson Reuters discontinues the indexation of retracted articles, although ALL previous citations done before retractions are retained. In other words, IF published in the Journal Citation Reports are never updated for retracted articles. Take the case of the 5 retracted JBC papers cited in this post. They accumulated 534 citations, which represent ca. 3% of citations received by JBC over one year. These citations, based on falsified/fabricated data, make obviously no sense once these papers are retracted, and the IF for JBC should thus be downgraded accordingly.
      I suspect that high impact journals would have significantly lower IF after correction for retracted papers. Basically because retracted papers are generally highly cited prior to retraction. I’m almost sure that the issue has been studied by others, but less sure if suitable figures have been released.
      Sorry for the long comment. Anyway I agree that publishers must publicize as soon as possible and in all media that a paper has been retracted, in order to avoid future unsuitable citations.

      • chris March 13, 2013 at 2:07 pm

        According to the ISI citation reports those 534 citations constitute around 0.13% (not 3%) of citations to JBC in the relevant year.

        I suspect removing citations to retracted papers would have a very minor effect on impact factors.

        In any case there may be very good reasons for citing a retracted paper. I do agree though that it’s important that it is made very apparent that a paper has been retracted.

        • Toby March 14, 2013 at 8:01 am

          Why should there be a reason to cite a retracted paper? Once retracted the data, interpretations etc. retracted papers become irrelevant, they disappear from the face of the earth, they do not constitute information, they are a dead parrot! So unless it is for the purpose of pouring scorn on the authors (not a worthwhile pursuit in my view) they are best forgotten.

          • chris March 14, 2013 at 2:20 pm

            There are several reasons why it might be appropriate to cite a retracted paper:

            A recent example on RetractionWatch is a good reason (i.e. because the authors are describing experiments made explicitly to address the apparently false findings in a retracted paper):

            http://www.retractionwatch.com/2013/02/15/mit-lab-retracts-cell-synapse-tagging-paper-for-falsification-or-fabrication/#more-12397

            Other reasons may be suggested by imaginary sentences in papers that cite retracted work; e.g.:

            “Although the paper describing a study on the effects of X has been retracted, an interpretation of the authors warrants further analysis in our opinion. We have addressed the implications of this interpretation in the context of Y and report this analysis here.”

            “Blogs et al. has been retracted. However we believe that their strategy for obtaining X from Y cells has some promise in the context of Z. We have attempted to assess whether some of the methods used by Blogs et al. might in fact be useful for Z and indeed find that with some optimization the broad strategy reported by Blogs et al can be effective.”

            “The results of the experiments described here lead to an intriguing idea namely X. We should point out that we are not the first to report this idea. Blogs et al. reported the same idea, although their paper was subsequently retracted.”

            etc.

            The fact is that a retracted paper doesn’t “disappear from the face of the earth”. It is still there in the print version in libraries and in accessible electronic archives (journal websites mostly), in the latter case hopefully prominantly marked as being retracted.

        • Sylvain Bernès March 14, 2013 at 10:15 pm

          @Chris. Regarding the negative impact factor of retracted papers, I did the following rough estimation:
          1- The five Smart’s (sic) papers retracted in the March 1, 2013, JBC issue have been cited 534 times.
          2- Raw data for JBC are: cites in 2011 to items published in 2010: 17937; cites in 2011 to items published in 2009: 19786.
          3- Merging 2009 and 2010 data affords ca. 19000 citations within a single year for JBC.
          4- (534/19000)*100 = 2.8 %
          Certainly, such a computation is far to be satisfactory. In order to get your 0.13%, you probably considered the full time gap for the publication of the five papers (1998-2005). Your calculated correction is correct within the gap.
          The point here is not to discuss what is the best metric for the correction of IFs accounting for retracted papers. The point is to make clear that IFs are necessarily overestimated because Thomson Reuters is handles static metrics rather than dynamic metrics. On the other hand, I think that high IF journals have higher overestimation than low IF journals (although I have no proof for that point).
          I think it’s somewhat ridiculous to release IFs with 3 decimal places, or to pop a bottle of Champagne because the IF of the journal in which you publish your works moved from 4.675 to 4.685. This move is nothing compared to what happened to Acta Crystallographica A, with a jump from 2.051 to 49.926 between 2008 and 2009, followed by a dramatic drop from 54.333 to 2.076 for 2010-2011.

  • ferniglab March 12, 2013 at 5:33 pm

    So, as ever for all attempts at using publication metrics as a proxy for quality: caveat emptor. Nothing beats actually reading the paper!

  • Post a comment

    Threaded commenting powered by interconnect/it code.