We’ve seen vigorous debates here on Retraction Watch about when studies should be retracted. Does it require fraud? Just not being reproducible? Somewhere in between?
Given the apparent divergence of opinions on the issue, we thought it would be worth highlighting a case that involves language we haven’t seen before. Here’s the notice for “Apoptosis of CT26 colorectal cancer cells induced by Clostridium difficile toxin A stimulates potent anti-tumor immunity,” which originally appeared online in Biochemical and Biophysical Research Communications in April:
This article has been retracted: please see Elsevier Policy on Article Withdrawal (http://www.elsevier.com/locate/withdrawalpolicy).
This article has been retracted at the request of the Authors.
After extensive discussions with the lead author and several coworkers, the authors concluded that several experimental results, especially results regarding toxin-induced apoptosis and dendritic cell maturation, need further repeats before the authors can make final conclusions. The authors therefore believe that the article published in its current form is too preliminary and potentially misleading to BBRC readers in this field and would like to apologize for this for this error.
We asked the corresponding author, Jufang Wang, of South China University of Technology in Guangzhou, and the editor of the journal for more details, particularly on what made the authors realize the results were too preliminary. We’ll update with anything we hear back.
In the meantime, this seems like a group of researchers is bending over backward to correct the literature — something some Retraction Watch readers may find objectionable, if previous comments are any indication. And we have to ask, how did it get published in the first place?
Another group has retracted a paper from the same journal for much more typical reasons. Here’s the notice for “Overexpression of αCP2, a translational repressor of GAP-43, inhibited axon outgrowth during development in Xenopus laevis,” originally published in February:
This article has been retracted: please see Elsevier Policy on Article Withdrawal (http://www.elsevier.com/locate/withdrawalpolicy).
This article has been retracted at the request of the Authors.
The authors have identified severe contaminations in their enzyme and protein products. After carefully reexamining their paper they have found some of their results were not reproducible and would like to retract their paper.
Hat tip: Clare Francis
Typical of the POP atmosphere! How many syntheses reported in, say, Tetrahedron Letters, have been carried out ONE time (with, of course, a yield including two decimal places…)? How many DFT optimizations are computed using a SINGLE functional and ONE basis set? And how many submission drafts are read by a SINGLE author, while the other co-authors just check if their name and affiliation(s) are OK?
At one time, that was the whole point of BBRC. It was originally delivered as a set of photocopies of typescript with a minimal cover glued on. The very point was to present preliminary results and avoid long publication delays. No one expected to be able to rely on those papers to the same extent as something printed in a glossy journal. Its still jarring to think of something being “too preliminary” for BBRC. True, in those days we carved Ehrelenmeyer flasks out of mammoth bone using flint knives. Things have changed. But even now, BBRC shouldn’t be held to too high a standard, or it will lose its ability to serve its mission.
“too preliminary and potentially misleading”
The authors have not conducted any new experiments to conclude that the published results are misleading. It looks like that they have found something by talking to each other and they want to retract the paper before somebody else writes to the editor.
”In the meantime, this seems like a group of researchers is bending over backward to correct the literature — something some Retraction Watch readers may find objectionable, if previous comments are any indication.”
That’s disingenuous. I haven’t noticed anyone arguing against the flagging of work that authors themselves have identified as being incorrect, and having the relevant paper(s) retracted. That’s the case with the Craig Hill retractions of papers on transition metal oxo complexes, the article in BBRC on this particular thread , the microRNA paper retracted from Molecular Cell when the authors discovered an artefact due to miRNA differential extraction using a commercial reagent, and so on.
That’s normal. A serious scientist that discovers a fatal flaw in their work will inform the wider community and retract the relevant paper if necessary. There’s lots of examples of this, and I haven’t seen anyone on these threads arguing against this practice. And retracting work that authors lose faith in isn’t “bending over backward to correct the literature”. It’s making a completely straightforward correction because serious scientists prefer to get stuff right; apart from anything else, if you cock up – you’d rather identify the cock-up than have someone else uncover it.
This is entirely different from other types of examples raised on Retraction Watch where, for example, it has been suggested that old papers in the scientific literature that have subsequently been shown to be incorrect should be hunted down and retracted (e.g. the dumb example of Linus Pauling’s incorrect DNA structure paper), or that authors might be bullied into retracting papers or making proscribed corrections because someone else happens not to like the work (the paper on FUND implementation in Ecological Economics). I suspect these are the type of examples that “some Retraction Watch readers find objectionable”.
It’s pretty obvious that there are different categories of erroneous publications and that there are differences in how these should be dealt with. As a rule of thumb no one should be forced into a retraction unless fraud has been identified, and anyone has the right to retract their own work.
Thanks as always for the comment, Chris, and for drawing these distinctions.
In response to your comment about that sentence being “disingenous,” however, I wanted to point to three recent comments that do in fact “[argue] against the flagging of work that authors themselves have identified as being incorrect, and having the relevant paper(s) retracted.” They’re all on one of the posts you mention, about the Molecular Cell retraction:
http://www.retractionwatch.com/2012/07/13/group-retracts-microrna-paper-after-realizing-reagent-was-skewing-results/#comment-16370
http://www.retractionwatch.com/2012/07/13/group-retracts-microrna-paper-after-realizing-reagent-was-skewing-results/#comment-16530
http://www.retractionwatch.com/2012/07/13/group-retracts-microrna-paper-after-realizing-reagent-was-skewing-results/#comment-16533
The BBRC example is pretty dodgy ‘though. It’s difficult to imagine a scenario in which someone publishes a paper (even a rush job, for example to stake priority, which is largely what BBRC is for) and then decides that actually some further repeats are necessary. You might imagine a situation where having rushed your result into print to stake your claim to priority, you then do a few more of the “just-to-make-sure” experiments you should have done before publishing anyway. If the results come out O.K. you sigh with relief and get on with the next bit of the work (no comment/retraction necessary); if the results look problematic then you correct or retract the paper for that reason.
And while we’re in a biophysical mood, here’s one in the latest issue of the Biophysical Journal:
http://www.sciencedirect.com/science/article/piiS0006349512007229
Biophysical Journal, Volume 103, Issue 2, 18 July 2012, Page 374
Jin Wang, Bo Huang, Xuefeng Xia, Zhirong Sun,
RETRACTED: Funneled Landscape Leads to Robustness of Cellular Networks: MAPK Signal Transduction
Biophysical Journal, Volume 91, Issue 5, 1 September 2006, Pages L54-L56
This article has been retracted at the request of the Editor-in-Chief.
The editors have noted that there is a substantial overlap of text and content between this Biophysical Journal article and the following article: Wang, J., Huang, B., Xia, X., and Sun, Z., Funneled landscape leads to robustness of cell networks: yeast cell cycle. PLoS Comput. Biol., 2 (2006) e147, http://dx.doi.org/10.1371/journal.pcbi.0020147.
etc. etc.
I don’t think there ever was a good reason for BBRC to exist but these days when journals judge papers within a month, there is even less. Of course scientists are competitive like most human beings and want to get their work published first. If they are doing original work, they dont need BBRC and if they aren’t, who cares?
On the question of whether papers with erroneous conclusions should be retracted, I believe that the best any of us can do is to get the data reproducible and the methods correctly used. That the conclusions may prove wrong, is inevitable.
Elaine Newman
A review which is delayed over one month is certainly seen as an offense by most chemists. I don’t know if the situation is worst in biophysics, molecular biology, and related fields. Regarding the couple of BBRC mentioned in the post, the first was published 24 days after being submitted, and the other one within 10 days… If the normal Elsevier process was followed, that means that in a 10-days window, the submission was transferred to reviewers, reviewed, the paper was revised, returned to BBRC with corrections, edited, and final proofs checked and approved by authors! All that work for an eventually retracted paper! In the same way, I remember the case of a non-Elsevier journal which, 10 years ago, was proud to “guarantee a 5-days review for all received papers”. I unfortunately don’t remember the name of that journal, which probably disappeared in the limbo of huge-editorial-board-but-never-read journals.