Retractions in Science
What happens when a mistake is found in a published research article? Everyday Einstein takes a look at scientific retractions and corrections.
Page 2 of 2
The reviewers will make comments and recommendations about the fate of the paper and send those back to the editor. The editor will then decide whether to accept the paper, reject it, or request additional information from the scientist. Collectively, this process is called “peer-review,” because the paper is being reviewed by the scientist’s peers.
Once a paper has been published, it appears on the journal’s website, is (usually) printed in their magazine, and is indexed in several 3rd party indexes, such as Google Scholar and PubMed. Other researchers can then find those papers when they are conducting their own research in similar fields.
Unfortunately the process is far from foolproof, and sometimes mistakes slip through the system. When a paper with a mistake is published, one of three things can happen. Either nobody will ever know, someone will realize there’s a mistake and will issue a correction (technically called a corrigendum), or the paper will be retracted, meaning it is stricken from the scientific record.
According to the research journal Nature, the decision to issue a correction versus a retraction comes down to the nature of the mistake and its impact on the overall message of the paper. If the core message of the paper is still valid, but there was some flaw in the methodology or presentation of the results, a correction is issued. If the core message of the paper is no longer valid, then the paper is retracted.
Unfortunately while the media typically reports wildly (and sometimes inaccurately) when a paper is first issued, almost nobody hears about a retraction (much less a correction) unless it’s associated with a sensational scandal.
Since many people store papers related to their field of research in their own personal reference libraries, they might never know that a paper they’re using to guide their research is flawed. Even some major 3rd party indexes suffer from this problem.
A 2012 study in the Journal of the Medical Library Association searched the internet for over 1,700 papers that had been retracted between the years of 1973 and 2010. Over 280 of those papers were still available in a variety of locations, including educational websites, major 3rd party indexing services, and commercial websites. And over 1,300 of the retracted papers were found in multiple online personal reference libraries.
This means that despite the fact that those papers have been retracted either by the journals or the scientists themselves due to inaccuracies, flaws, or flat out scientific misconduct, those papers still sit in the public domain waiting to be used as the rationale behind further research, medical decisions, and public policy decisions.
Who Watches the Watchers?
Fortunately if you’re interested in keeping up with which bits of science turn out to be not as scientific as originally thought, there’s a fantastic resource you can turn to called Retraction Watch.
Retraction Watch is a blog started in 2010 by Ivan Oransky and Adam Marcus to address several issues they found with the retraction process, including the lack of transparency that often masks why a retraction occurred.
If you’re interested in the research presented in a paper that was retracted, it’s important to know just why that retraction occurred. Was it because of technical errors which led to the wrong conclusion? If so, you might be able to deduce a different, though similarly interesting conclusion from the same paper. Or was it because the scientists lied about the experiments ever having been done in the first place? Retraction Watch seeks to answer these and further questions.
So now you know more about scientific research papers and retractions, as well as how to keep up with what has been retracted and why.
If you have a question you’d like to see on a future episode (which I promise to try and read more carefully than I did that windshield wiper question), send me a message at email@example.com, or on twitter at @QDTEinstein.