Dear Irene,

I review many manuscripts every year so that I will answer you after my experience that could be different from others. I always start from the idea that each author is an honorable and honest colleague that is reporting on his/her research. My duty as a reviewer is not to judge the personality of an author, but the content of his/her manuscript. In this context, I have reviewed excellent new original ideas, other less exciting, and others very poor but I have never had the "feeling" that an author has falsified data. As a reviewer my duty is to assist an editor/journal in the evaluation of the content of a specific paper. Anyhow, if I could have a doubt, I would transmit it confidentially to the editor, but I would establish that this is only a doubt. We cannot accuse anyone on a suspicious, a suspicious or doubt is only that. 

You ask what can we do to prevent a fraud. I believe that you are asking an impossible question. We cannot distrust each colleague as a potential criminal. We have to trust also the high principles of Academia and ethic behavior and we should educate students on such principles, but we cannot go around distrusting others. If you compare the high number of publications/year and how many cases of fraud are discovered, certainly they are few. This gives us hope!

As an editor I have had other kind of experience, not that a reviewer is detecting a fraud, but an author has openly accused other of taking his/her ideas from a manuscript or during a presentation in a meeting, etc. But these are cases where you have to deal with specific and open accusations and I can tell you that those are very unpleasant experiences. 

With my best wishes,

Gloria

Your background information to this case is incorrect. The university (Uppsala University) did NOT conclude that there was a misconduct. On the contrary, UU dismissed the claim by the whistle-blowers that the study could not have been carried out in the manner claimed.  

However, now Sweden’s Central Ethical Review Board (CEPN) has come to the conclusion that there was misconduct, see http://www.sciencemag.org/news/2017/04/paper-about-how-microplastics-harm-fish-should-be-retracted-report-says

I find your questions a bit off the mark. This case has not been about spotting fraud by reading the published study. It's been about the realization by researchers who were at the field station when the study was supposedly carried out, that this could not be correct, and subsequently, that the authors could neither provide the original data or any further proof of the study actually having been carried out. 

The CEPN report also criticizes Science for having accepted the manuscript, as the necessary supplementary data were never provided 

My question was, whether a reviewer could detect a problem in the data by carefully examining the available information. The review process is supposed to act as a safegard against work, that does not meet the scientific standards.

Reviewers have to do more than read the abstract and look at the graphs - as most readers do unless they are interested in details of the method. Science provides a supplement to the paper describing methods and some data, so there was more information for reviewers than the publication itself (which actually looks very nice).

So I repeat my question: would you have spotted the fraud?

If the reviewers had little chance, the second question is, who should do what to prevent fraud? Journals? Editors rely on reviewers. The University?  The Supervisor?

Dear Irene,

I review many manuscripts every year so that I will answer you after my experience that could be different from others. I always start from the idea that each author is an honorable and honest colleague that is reporting on his/her research. My duty as a reviewer is not to judge the personality of an author, but the content of his/her manuscript. In this context, I have reviewed excellent new original ideas, other less exciting, and others very poor but I have never had the "feeling" that an author has falsified data. As a reviewer my duty is to assist an editor/journal in the evaluation of the content of a specific paper. Anyhow, if I could have a doubt, I would transmit it confidentially to the editor, but I would establish that this is only a doubt. We cannot accuse anyone on a suspicious, a suspicious or doubt is only that. 

You ask what can we do to prevent a fraud. I believe that you are asking an impossible question. We cannot distrust each colleague as a potential criminal. We have to trust also the high principles of Academia and ethic behavior and we should educate students on such principles, but we cannot go around distrusting others. If you compare the high number of publications/year and how many cases of fraud are discovered, certainly they are few. This gives us hope!

As an editor I have had other kind of experience, not that a reviewer is detecting a fraud, but an author has openly accused other of taking his/her ideas from a manuscript or during a presentation in a meeting, etc. But these are cases where you have to deal with specific and open accusations and I can tell you that those are very unpleasant experiences. 

With my best wishes,

Gloria

Dear Gloria,

Thank you so much for your detailed answer! Actually I was really interested, whether in this specific case reviewers would have had the chance to get suspicious about the data. I think nobody, who has to review a paper will immediately scream "fraud", but one might want to see more of the data. (Fortunately I never had to do that as a reviewer). The procedure you suggested for such a case seems perfectly fine to me.

The questions about who had which idea first and who borrowed it from whom can be difficult. I can imagine, that dealing with them must be extrodinarily annoying and I sympasise with you. However, this is not as damaging to the field as a whole as fraud.

Kind regards

Irene

Here is the summary of the issues with the paper: https://drive.google.com/file/d/0B6KTULK8SWmibWJpMGdMZlRzT0g4NmlhSG1kOFJaLUtfUUow/view

It seems like some of the methodological issues could possibly have been caught by a careful reviewer - like the issue of exaggerated sample sizes.

I am compiling a summary of reactions to my posts on researchgate and the fishlist.

@Björn, @Gloria, @Joacim, Thank you very much for your contributions!

@Björn: thank you for correcting my mistake about the University! I read their conclusions about the case and found them a bit .... - especially when you compare it to the documents, which are available from the links you and Joacim provided.