Back in August we — and others — wrote about a paper in Organometallics for which one member of the study team appeared to have instructed a co-author to pad the article with artificial results. From the supplemental information (SI) of that paper:
Emma, please insert NMR data here! where are they? and for this compound, just make up an elemental analysis…
Now comes a correction statement from the group that can only be described as “mega.” First reported last month by Chemical & Engineering News, the lengthy notice begins:
During peer review of the manuscript, a reviewer recommended and the assigned Associate Editor relayed a request that the manuscript be shortened. In response, the authors moved the Experimental Section provided in the submitted manuscript to a Supporting Information file. In the course of this text transfer, the authors regrettably introduced new text consisting of internal communication among the authors not intended for publication. These files were not provided to reviewers. The purpose of this Addition/Correction is to share further amendments to the published article and Supporting Information, providing additional experimental detail as warranted by further editorial review.
In the course of editorial review, microanalysis reports were provided to the Editor for all but six of the compounds reported. For the six compounds where original microanalysis reports were not located (compounds 4, 5b, 11a,b, 12, and 15b), and for all compounds in this article, 1H and 13C NMR spectra were provided and included in the appended Supporting Information in accordance with the Organometallics guidelines to authors.
The notice goes on to describe 19 specific changes to the article, and links to nine more in the supplemental information. Here are the first three:
(1) The following sentence beneath Scheme 1 is incorrect and should be deleted: Nevertheless, the elemental analysis matched that predicted for Pd{(M,SS,SS)-p-tolyl-binaso}Cl2 (4), and no benzonitrile was observed in the 1H NMR spectrum.
(2) The assigned stereochemistry for the X-ray structure of 6 is incorrect. The first sentence in the caption for Figure 2 should be corrected to read: Figure 2. Molecular structure of Pd((P,RS,RS)-p-tolyl-binaso)(TFA)2 ((P,RS,RS)-6·2Et2O) (50% probability ellipsoids; hydrogen atoms and solvate molecules have been omitted).
Likewise, the sentence before the crystallographic discussion of compound 6 beneath Figure 1 should read: Crystals of the complex (P,RS,RS)-6·2Et2O were grown, and an X-ray analysis was performed (Figure 2).
(3) The palladium complexes 4, 6, and 7 have been synthesized starting from either Pd(PhCN)2Cl2 or Pd(CH3CN)2Cl2.
This might not be the most extensive correction we’ve covered, but if not it has to be close. We’ll leave it to our readers to say whether a retraction might have been the better option here.
The correction is accompanied by a note from editor in chief John Gladysz and associate editor Lanny Liebeskind. Excerpt:
The editorial review is now complete. This has resulted, with the authors’ cooperation, in the publication of an Addition/Correction and a revised SI file in this issue.(2) The former summarizes the key changes. Of the 16 new compounds, original microanalysis reports (certified by University officials) were located for 10 and transmitted to the undersigned. There was no evidence in any of the materials received that indicated falsified analyses. Among other characterization data, NMR spectra for all 16 compounds are provided in the SI. The Author Guidelines for Organometallics(3) allow NMR spectra to be used in lieu of microanalyses as bulk purity criteria, although (as noted in the guidelines) this is not as rigorous a measure. These and other issues dealing with compound purity are regularly debated within the community and will be treated in an upcoming Editor’s Page.
The peer review undertaken for this manuscript adhered to the same rigorous standards and procedures as those followed for any submission to Organometallics. The embedded comment that raised questions was not present in the SI that accompanied the originally submitted manuscript. All reviewers returned thorough reports of the original submission, and one suggested that the manuscript be shortened prior to publication. In response, the authors moved the experimental section to the SI. The undersigned assumed that such a cut/paste revision did not necessitate a line-by-line reexamination of the manuscript and the accompanying SI, with the result that the embedded comment remained undetected until Chemical Abstracts indexing staff alerted ACS publications several weeks after the article appeared on the web. At that time, the undersigned immediately reached out to the corresponding author and began the process that ultimately resulted in the Addition/Correction and revised SI file noted above.
Organometallics takes all cases of alleged research improprieties very seriously. Input is always welcome from all sources. The journal follows long-established procedures for reviewing and taking appropriate actions, where warranted, to preserve the scientific integrity of the published literature. In addition, Organometallics has a complement of precautionary policies regarding submissions by authors whose past actions have not been in accordance with the ACS Ethical Guidelines to Publication of Chemical Research.(4) Institutions and funding agencies also have established departments, policies, and procedures for investigating allegations of ethical misconduct by researchers and are positioned to implement appropriate penalties or corrective measures when warranted.
It is hoped that this Editor’s Page and the referenced materials will help to address some of the concerns that have surrounded this article in recent months. The ultimate responsibility of the Editor in Chief for all problems that arise in the course of publishing this journal, and their consequences, is duly acknowledged.
I think it has clarity and is thorough… I prefer a mega-correction with an explicit note by the editor, than an unhelpful “we’ve posted an updated SI file” notice.
I think that retracting the article and republishing the corrected article at the same time would create a final product that would be much easier for fellow researchers to follow and interpret. However, it would likely be perceived as an even bigger black mark on the authors. Best solution is to get it right the first time, I guess!
“In the course of editorial review, microanalysis reports were provided to the Editor for all but six of the compounds reported. ”
But that’s not the point.
The original suspicious text was “and for this compound, just make up an elemental analysis…”
The question is whether the authors actually made up an analysis for publications, or could they produce the original document.
I was one of the defenders of this group and at the time I said that the unfortunate text could have been an internal joke. Bu the issue is not whether the group can eventually produce data to support their paper and claims. The issue is whether they were making up data to fraudulently claim work that they had not done. This lengthy notice doesn’t seem to address the main point of concern.
However, a very positive consequence of the revised SI is the account of elemental analyses in the correct format: how many analyses were done is specified, and ALL figures are quoted, along with averages. For example:
“Calculated for PdC68H52B2F8O4S4.3H2O: C = 58.53%, H = 4.18%; Found (triplicate analysis, identical sample): C = 58.44%, 58.61%, 58.65% (avg, 58.57%), H = 3.91%, 3.98%, 3.95% (avg, 3.95%).”
This makes a big difference with useless numbers generally given in the experimental section, like this one, taken from the SI of the article following the Dorta’s paper in the February issue of Organometallics:
“Anal. Calc. for C56H42IrN5 (%): C, 68.83; H, 4.33; N, 7.17. Found: C, 70.22; H, 4.47; N, 7.36.”
Such numbers are not helpful for assessing sample homogeneity; nothing is commented regarding the observed deviation for C content, which would be worrying if a triplicate analysis was carried-out, etc.
SI sections should not be considered as a trash for numbers, poorly scanned spectra and experimental procedures deposited here with the hope that nobody will attempt to reproduce them … It’s an essential piece which add nothing to the paper but rather allows to assess the correctness of the work.
Now, if private jokes are inserted in the preliminary drafts of a SI section, that’s fine for me. That may reflect a good atmosphere in the Lab. But check (twice) the last version! T-W-I-C-E !!!
”I was one of the defenders of this group and at the time I said that the unfortunate text could have been an internal joke”
Or they were using ‘to make up’ in the sense ”To supply the shortfall or deficiencies of’ or ‘To bring up to a given number’ [Oxford English Dictionary; http://www.oed.com; accessed 5 March 2014].
“However, a very positive consequence of the revised SI is the account of elemental analyses in the correct format:”
But the point is whether these numbers were made up!
I can publish some very impressive statistics, if you accept that it is OK to make things up.
I wish Springer would do the same for these two papers, which have, I estimate, about 50 (PCTOC) and 80 (APP) errors (minor, questionable, or large) in them:
Plant Cell, Tissue and Organ Culture (PCTOC) March 2013, Volume 112, Issue 3, pp 361-368
Primary and secondary somatic embryogenesis in Chrysanthemum cv. Euro
Aung Htay Naing, Chang Kil Kim, Baek Ji Yun, Jo Yu Jin, Ki Byung Lim
http://link.springer.com/article/10.1007/s11240-012-0243-5
Acta Physiologiae Plantarum October 2013, Volume 35, Issue 10, pp 2965-2974
Primary and secondary somatic embryogenesis in Chrysanthemum (Chrysanthemum morifolium) cv. ‘Baeksun’ and assessment of ploidy stability of somatic embryogenesis process by flow cytometry
Aung Htay Naing, Jeon Su Min, Kyung Il Park, Mi Young Chung, Sun Hyung Lim, Ki Byung Lim, Chang Kil Kim
http://link.springer.com/article/10.1007/s11738-013-1328-4
Despite detailed lists of the errors to the entire editor boards of these two journals and the Springer management, only these two pathetic, unrepresentative, errata:
http://link.springer.com/article/10.1007/s11240-014-0427-2
http://link.springer.com/article/10.1007/s11738-014-1501-4
Shame on Springer and shame on these editor boards. They bring disrepute to the plant sciences for not taking firm and decisive action.
Dan, I understand your point, and I agree that make up raw data is the most offending misconduct in chemistry and related sciences. However, as detailed in the note issued by Editor and associate Editor, a post-ASAP review has been carried-out, which concluded:
“There was no evidence in any of the materials received that indicated falsified analyses”.
This is exactly what you asked for. Of course, you may have your own reasonable doubts about that review. Sincerely, I think we must acknowledge the work done by Organometallics, which was probably very time-consuming, and trust in final conclusions. Otherwise, I will start to question numerical values of fundamental physical constants published by the NIST.
you also have to be careful because “to make up” in the sense of “to falsify” is a somewhat vernacular english usage; and foreign speakers of english might use “to make up” in the sense of “to make”. And things can get muddier; a nonnative speaker of english in a professorial role might misuse the word “make up” in the sense of a “makeup” test – something added to enable the completion of a (grade or publication) in lieu of its being missed earlier.
The professor is a big wheel and in the Anglosphere for most of his career. And the grad student had moved on to another position in another country away from her samples.
He asked her to make up the sample and hasn’t done one thing to deny it. Bad, bad news. I just wonder if this was the first time he pushed for this sort of thing or if he did it a lot. The inflated yield and solvent when none was there is exactly the sort of padding I expect from a corner cutter.
1. What the journal did after seeing the remark was (in analogy) to ask for the lab notebook and check the actual work. This is a way more serious inspection than normal and was done because of the remark advising fabrication. The result of that inspection showed OK analyses for some compounds, but for several others LACK of evidence. Thus removal of original elemental analysis (I think) and reliance on NMR results. This is probably OK to validate the basic findings of the paper.
2. Note that even though the basic paper wrt molecules is OK, the scrutiny led to several detailed analyses (of elemental analysis) being retracted. [I think.]
3. The journal leaves it unsaid the moral issues and has actually said that it is not their job to take action for the original moral failing. This is the responsibility of the institution and grant agency. IMO, an investigation should be done and the professor should be compelled to apologize for his remark. [That he has not done so already, proactively shows poor character.] Earlier work should also be investigated and if there is falsification found, he should be terminated.
4. The sheer complexity of the correction is too large to allow for easy reading of the result. Instead, a retraction and corrected version should have been done. With re-review. The retraction could have happened before the corrected version went through. The way this was handled was over-kind to the authors and not kind enough to the readers.
This is the damning part: ” For the six compounds where original microanalysis reports were not located (compounds 4, 5b, 11a,b, 12, and 15b)….”
IOW, they couldn’t back up 6 claimed elemental analyses. That is bad, bad news after the earlier remark to do falsify data. The linked C&E article (from the same association that publishes Organometallics) goes way too far in saying that the investigation cleared up any issues of fabrication.
FWIW, I would like to know if the provided elemental analysis reports were all dated from before the paper submission or any done afterwards, any changes made to values. OK! I just read it. If you look at correction number 8, many of the compounds had replicate analyses done. IOW, it sounds like they couldn’t find those analyses either and redid the work.
SEVERAL mistakes in stereochemistry, diagrams etc. are apparent in the numbered corrections. IOW, the “molecule results” WERE CHANGED.
Corrections 10 and 16 are damning also and show a noted area where chemists fudge their results to make themselves look better. For 10, they claimed a solvent when it wasn’t. For 16, the yield was revised down. In both cases, the original results made them look better and the new ones make them look not so adroit. It’s like someone making a mistake in counting change…but always in one direction (to help himself). If you talk to people like Chembark, etc. they will clarify that these are long known issues/suspicions. Can find discussions pre-dating this paper covering exactly these problems (padded yields, invoking solvent).
Also, of interest, there were several comments made on the Chembark site’s multiple Dorta threads by a person linking to the ‘exoneration’. He or she did not engage in any discussion. And (IMO) misrepresented the mega-correction as an exoneration.
“you also have to be careful because “to make up” in the sense of “to falsify” is a somewhat vernacular english usage; and foreign speakers of english might use “to make up” in the sense of “to make”.”
That might be the case but 1) In context this interpretation is untenable; and 2) the authors have not made this claim.
As far as I am concerned, “just make up…” is unambiguous.
In answer to the question “Retraction or Mega correction” I think Mega Correction fudges the issues. The paper is clearly different and the scientific approach would have been to retract and then submit a new papers). A few have actually retracted when they found a SNAFU in their data and have gained considerable accolade from the community. Not retracting leaves a strong whiff of suspicion, regardless of the evidence, simply because the community is relying on third parties for the evidence.
I have yet to see a Mega Correction that would fall into the “Doing the right thing” category and it is probably medically inadvisable to hold one’s breath until such time….