Last week, we reported on a new paper by Scripps Research Institute researchers in which they described how two of their previous papers had been based on mistaken interpretations. The authors wrote in their new paper that they were retracting the earlier works, but the journal had told them the papers would be corrected instead.
We had asked Protein Science editor Brian Matthews for clarification, and he emailed us late last week:
When Dr. Kelly submitted this manuscript he explained that the new results had led him to the realization that the interpretation of the data presented in two earlier Protein Science articles was incorrect. The experiments described in the original reports were carried out in a technically satisfactory fashion and were reproducible. It was the interpretation of the experiments that was at fault. In his letter of submission Dr. Kelly indicated that he would be willing to either retract the earlier articles or handle the matter as corrigenda. His new manuscript was subject to standard scientific review and, after revision, judged to be acceptable for publication. The matter was also considered by legal staff at the publisher who advised that the correction should best be made via corrigenda statements. We anticipate that these statements, together with the above-referenced article, will appear in the November issue of Protein Science.
The article from Dr. Kelly which was posted on the Protein Science website was a working version which referred to “retractions” rather than “corrections”. It should not have been released and I apologize for the confusion.
So, the papers will be corrected, not retracted. The distinction drew a lot of comments on our original post, so we thought we’d poll Retraction Watch readers in another installment of “Ask Retraction Watch:”
[polldaddy poll=7362898]
There’s a difference between perceived deception and being “incorrect.”
Absolutely nothing should happen to papers later found to have ‘misinterpretations’. Science is built by interpreting results based on our knowledge at the time. Often, new results alter the way we interpret prior results. This is how science works. We continuously test new hypotheses. I would bet that if we go back and reread many papers from the past, based on our current knowledge, we would think that the conclusions drawn from the results were preposterous. Interpretations are always a consequence of our knowledge base at the time the data were interpreted. Thus, I do not think anything should be done with the prior publications. They should remain as is.
Lynn Wecker – I am not a scientist, I have not read the papers in question (if I did, I probably wouldn’t understand them), and I am curious about your response.
It seems to me that the relatively short time since the two papers were published (2009 and 2010) and Dr. Kelly’s desire to publish a correction are adequate justification to do so. A correction does not (necessarily) imply deception; it could just be a mistake. I assume that scientists rely, to some degree, on the interpretations of other scientists (not only results), and a public statement that one no longer stands by an early interpretation reflects the kind of openness and community spirit that I would hope most scientists would applaud.
I don’t think that this correction is mandatory, nor that all earlier misinterpretations, no matter how old, must be corrected. But I do think that this is a good thing for Dr. Kelly to do.
Is there something in your opinion that I am missing? Do you think that it would be WRONG to publish the correction, a waste of time, or a smear on Dr. Kelly’s reputation? I don’t understand why you object, and I would like to understand.
I am not trying to challenge you; I am genuinely interested in understanding your stance.
Respectfully,
Ken Pimple
Dear Ken Pimple,
I totally agree with Lynn Wecker. The corrections are a total waste of time. Given the accuracy of the data, it is obvious that the author -or other authors too! – can give a different, even better, interpretation. Let’s think about ptolemic interpretations of planet movements…
If the authors want to correct (its their work and interpretation), the (independent) referees agree and the Editor also, why complain? It might set a precedent, but not to the extent that all papers where some of the interpretation is wrong will have to be corrected I would think…
It’s not that I object to anything, nor do I think that a correction is wrong by any stretch of the imagination. Further, I do not think that the correction would be a smear on Dr. Kelly’s reputation. I just do not think it is necessary.
There was no deception. However, you imply that there could have been a ‘mistake.’ This is the problem. Perhaps, there was no deception nor mistake. The data were interpreted based on the current state of knowledge. With new evidence and knowledge, now the data can be interpreted differently.
In my mind a correction implies that something was in error, which, in fact, may not have been the case at the time the paper was written. Science continues to evolve based on new information.
So I am not objecting to anything. I just do not think it is necessary.
Thank you for the clarification. I appreciate your taking the time.
Best wishes,
Ken
Well said.
1. Thomas Kuhn et. al.
2. Perhaps I am too old — or perhaps it was not too long ago — but there was a time when people actually published subsequent erratas. Famous people. That is a proper and humble way to deal with errors.
This is the sad side of the whole retraction business.
More poll options needed – there should always be a “do nothing”!
A misinterpretation has quite a different character to a technical error, and debates over interpretation are an important part of the scientific process. Bad interpretations should not be buried in the same way as bad data.
What I would want to see in such a case is the original article preserved, with an addendum explaining the error and giving the new interpretation.
well said!
One serious problem I have encountered recently is that, following post-publication peer review, even after making an objective critique of papers and presenting this list of errors and scientific faults to the editorial board and publisher, that not even an erratum is being published. Even if we give the authors the benefit of the doubt (or the editors, depending on your perspective), why are publishers being so reticent about publishing errata when the peer pool critiques a paper post-publication? Publishers (and their team of editors and peer reviewers) should be held to the same level of critique as authors, and should also suffer consequences if they have found to overlook quality control during peer review. No system is infallible, especially since all peer review relies on humans (and their weaknesses), but then transparency and accountability must be placed first because there is no true way of differentiating deception from human failure.
Yes. I too feel that the changing landscape of publication is the key factor when one tries to understand the current state of retractions, particularly these somewhat peculiar cases.
In above comment I referred to the good old days (sic!) during which retractions were practically impossible. Paper copies were distributed all over the globe and hence errata was the only sensible option. The fact that you can now take a paper offline based on “false interpretations” is sad indeed. How about: “some of my early papers were weak and so maybe I just retract them?” The possibility alone shows that there is something seriously wrong in the current state of publication. Weird because the surrounding Internet goes precisely to the opposite direction, leaving traces and copies of anything you put online.
As for the solutions, journals should maybe reserve sections for this kind of stuff. Many good ones already have sections for debates etc. And actually, the fact that we have to write here, is part of the problem: these questions belong to the journals!
Everything is misinterpreted to some extent – has to be. So no problems. They could correct or at least the correction could point to the new paper, but certainly no more. After all, the new interpretation cannot be absolute. This is science, not metaphysics.
Time to come up with a new interpretation in science? Anything from centuries to a day, so no there should not be an issue with the 3 years either.
Of course, no retraction (retraction shows that the paper should not have been published at the time). But, of course, a section of the journal should occasionally appear as “Our Past Publications”, where the authors can publish explanations. In all cases, let people talk! And, if the author wishes so, give reference to this later explanation on the original e-article.
Not sure I agree with these comments. This wasn’t just a case of misinterpreting data, the data itself is flawed. I think in this case a retraction would have been the correct course of action.
The way to get these papers retracted is to approach the FULL editorial board and make sure that your logic is crystal clear as to why the papers deserve to be retracted. I can assure you that if you are right, truly right, no editor wants to look the fool in front of his peers for doing the wrong thing. Time to pressure the members of the editor boards to take action, especially when the publishers just sit on the side-lines in their hammocks. However, be prepared to be marginalized, called a cyber-bully or worse, and discriminated against for speaking the honest and brutal truth.