11

If a researcher publishes a paper found to be fraudulent, are the reviewers and editors of the paper also discredited for having approved the fraudulent paper?

22

A note about the nature of academic fraud: Merely being wrong is not fraud.

In the case where something is indeed fraud, it will depend very much on the paper in question - it would only discredit the reviewers and authors if there is the reasonable expectation that they could catch it.

Consider some examples:

Fraud by Omission: You run a study twenty times, and report the one version that gave you a significant outcome. That study is well-conducted and correctly analyzed - it just also happens to be cherry picked. How could a reviewer know that?

Fraud by Modification: You alter your data in some way to give you a desired answer, and then re-analyze it. This analysis is done correctly, you provide the reviewers with this "raw" data and they reach the same conclusions, etc. There are some mistakes you might make that could alarm a very alert reviewer, but it's also possible, being aware of these mistakes, that you wouldn't make them.

Similarly, if you edit some figures, reviewers might catch egregious versions, but some people are caught out by things like particular artifacts in the image files, which is not something to necessarily expect the reviewers to catch, or even necessarily to have seen (as sometimes we don't get the raw figures, but versions processed by the editorial software).

Fraud by Just Outright Lying, Absurdly Unethical Studies, etc. Most of the vaccine denial literature falls in this category. Deliberately misleading statistics, results not supported by the data, etc. These I do expect reviewers to catch, and do think less of the journals that let them go into print.

  • 4
    There must be "willful intent" for it to be fraud, otherwise it is just poor scholarship. – TheDoctor Oct 28 '17 at 4:18
9

The general answer is NO.

Academic peer review is not designed to catch fraud. The purpose is to review methodology, check assumptions and reasoning, and verify that the author's conclusions are justified given the data and analysis presented. While it's possible that a peer reviewer might notice some discrepancy and follow up, it's not expected and it's not likely.

Peer reviewers are domain-specific experts, they're not detectives. They do not have the time or the training to fact-check submissions beyond their own expertise and ferret out lies. It may in fact be impossible for a peer reviewer to prove fraudulent activity because most papers do not include their full raw data and many peer review systems are double-blind reviews. Even if you know the authors, a peer reviewer does not have subpoena powers and cannot compel evidence like a court can (the traditional venue for prosecuting fraud).

(As an aside, this is why replication, public datasets, and open-source code is viewed as highly favorable by many academics. Peer review will not catch frauds, but replication studies will fail to reproduce fraudulent findings. Public datasets and open-source code are their own form of proof that a paper's findings are correct.)

There are exceptions to the above. For example, there are rings of editors and reviewers who publish each others' work with little review or oversight as a form of quid-pro-quo and there are pay-for-publish schemes. In such situations where it can be shown that there is group wrongdoing then of course all the members of the group will pay a price.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.