Suspected research misconduct and poor research practices

First, PLEASE don't accuse any authors of misconduct on PubPeer. We'll explain why in this article.

Firstly, we are scientists. We should only work with data and logic. Our conclusions must be verifiable. On PubPeer, that means that the facts must be restricted to those accessible to other readers. The published record, in other words. (Unfortunately, we only rarely have access to original data). What none of us can verify is any conclusion regarding precisely how or why an apparent instance of misconduct occurred. In particular, the state of mind or the intention of a researcher is not a verifiable fact. Comments based upon personal knowledge or hearsay are unacceptable for the same reason - you CAN'T say "My friend used to work in the lab and said they cheated all the time". If you do have serious concerns along these lines, you might however consider contacting some of the investigative blogs (for instance Retraction Watch, who seem to be very adept at examining the collected works of suspect researchers) or raising your concerns with the relevant authorities, although we understand that this can appear risky and that they are often unsympathetic.

At the time of writing, PubPeer is receiving many comments of the sort pioneered by the hardworking members and followers of the Retraction Watch blog ( ). These often indicate odd coincidences in or poor preparation of image figures. Typical examples might include duplicate images, gels with strange similarities, etc.

Some problems with images are obviously more serious (e.g. montage of individual gel bands) than others (poor contrast). However, it is perfectly valid to point out all issues, IN A FACTUAL MANNER. Thus, it is acceptable to state that "band X appears to be surrounded by a rectangle with different background to the rest of the gel". It is NOT acceptable to state that "The authors have deliberately pasted in a different band". For those interested in guidelines for the preparation of image data for publication, and the kinds of problems that can arise, please see, for example, the current guidelines at Nature and an informative slide show by a Nature Editor ( ). There are many other resources, some of which are referenced at the end of the slide show

The second reason for sticking to publicly verifiable facts and a description thereof is of course legal. It is a foregone conclusion that PubPeer will one day be sued for libel or defamation by a disgruntled senior researcher. Even baseless law suits can be very costly to defend and the outcome is never certain. PubPeer may not have the financial resources withstand a law suit. In many jurisdictions, if a statement is made along the lines of "X deliberately falsified the data", we would be in the position of having to prove each step of the falsification and also the state of mind of the researcher (that it was done deliberately). The standard of proof can be very exacting and require information to which we would not have access (especially the private thoughts of the researcher!). The British libel system is notorious in regard of the difficulty and cost of defending even totally meritless suits, but lawyers cost a lot of money everywhere!

If, however, (implicit) criticism on PubPeer is couched in terms of simple factual descriptions of published work, it should be a straightforward matter to establish the veracity of those facts and to dissuade the legal attack entirely or at least to get it thrown out at an early stage.

Moreover, your readers are mostly highly intelligent scientists. There is absolutely no need to spell out any suspicions you may have about the authors' state of mind or intentions, however righteous. You readers will be able to draw their own conclusion without the slightest difficulty.

Note that not only will PubPeer be on the hook for critical comments. You could also find yourself a defendant. It is quite likely that we would be forced to divulge the contents of our database and server logs in a law suit. So let's stick to the facts.

In order to keep discussion factual and minimise legal risks for everybody, we reserve the right to remove or edit comments that do not conform to these guidelines or in our judgement expose us and you to legal risk in other ways. Please help us by "reporting" (a button is provided on each comment) any comments you believe might expose us to legal risk.

2017-03-09 10:28:28 2017-03-09 10:28:28 7 howto How to comment on PubPeer


We explain the rules for posting on PubPeer and offer some advice on writing effective comments.



By far the most important rule for commenting is to base all of your statements on *publicly verifiable* information. This will usually be the data published in the paper you are commenting on, but could also be another paper or some other source such as a book, newspaper or web site. Please cite your sources to allow easy verification.

Observation of this rule automatically focuses discussion on the facts and the science. It also greatly reduces the legal risks to PubPeer and to you. If comments only involve immediately verifiable facts there should be no difficulty in proving the truth of the statements. The truth is of course a strong defence in any defamation/libel case. However, our aim is not just to have a strong defence, but never to need one. If your comment only contains the obvious truth, that should discourage unhappy authors from initiating any legal attack.

Concentrating on the published facts means avoiding any speculation about the intentions or actions of authors. Although it is almost inconceivable that some of the anomolies highlighted on PubPeer arose in any innocent manner, accusations of fraud, misconduct, deceit or incompetence are nevertheless forbidden on PubPeer. In any case, a clear factual argument is usually all that is required to allow readers to draw their own conclusions, as you will observe as you read the comments. See also .



 Be polite; this should follow fairly automatically from being factual.

 Although it is important to be concise, a minimum of context and explanation are very helpful to readers, so please consider providing these even if that does require extra work.

 Please also proof-read your comment carefully before posting it. Registered users can edit or even delete posts until somebody else replies, but comments made via the unregistered route obviously cannot be altered once submitted.

2017-03-09 10:28:28 2017-03-09 10:28:28