We’ve written before on scientific fraud and the problem of how easy it is to get papers with fake or manipulated data published. These “studies” that somehow make it through the publishing process can range from relatively harmless, such as the deliberately faked “chocolate is good for weight loss” study, or they can have major detrimental effects, like Andrew Wakefield’s now-discredited MMR vaccine/autism paper, which essentially sparked the modern anti-vaccination movement. (The journal that published it, The Lancet, later retracted the study and Wakefield lost his medical license, but the damage was already done).
An article at Discover blogs titled “The Perfect Scientific Crime?” highlights the relatively simple steps involved in getting fake data published without being caught. The author writes, “This post is not my practical advice for fraudsters. Rather, I am trying to suggest fraud-prevention tips.”
First and foremost, suspicious data are usually what gets a fraudster caught. Either the numbers themselves are suspicious – by being too “neat” or “too good to be true” – or the method of data collection could not have happened as described. Suspicious data then lead to an investigation.
Also, because publishing a single-author study would attraction suspicion, a perfect fraudulent paper would need coauthors. But finding a coauthor to help publish a fake study would likely be a difficult endeavor. By putting their names in a paper, coauthors are sharing responsibility and thus vouching for the validity of the study – which makes scientific collaboration on a study a good first step to prevent fraud.
Additionally, requiring authors to publish the raw data alongside the summary results would be another useful prevention method, because it is more difficult to fake raw data than statistical summaries of data.
What this article didn’t mention was the loose policies surrounding peer review – for example, authors being allowed to choose who they want to peer review their studies. Stricter checks and balances in peer review need to be more firmly established to catch fraudulent studies before they even get published.
An article at Discover blogs titled “The Perfect Scientific Crime?” highlights the relatively simple steps involved in getting fake data published without being caught. The author writes, “This post is not my practical advice for fraudsters. Rather, I am trying to suggest fraud-prevention tips.”
First and foremost, suspicious data are usually what gets a fraudster caught. Either the numbers themselves are suspicious – by being too “neat” or “too good to be true” – or the method of data collection could not have happened as described. Suspicious data then lead to an investigation.
Also, because publishing a single-author study would attraction suspicion, a perfect fraudulent paper would need coauthors. But finding a coauthor to help publish a fake study would likely be a difficult endeavor. By putting their names in a paper, coauthors are sharing responsibility and thus vouching for the validity of the study – which makes scientific collaboration on a study a good first step to prevent fraud.
Additionally, requiring authors to publish the raw data alongside the summary results would be another useful prevention method, because it is more difficult to fake raw data than statistical summaries of data.
What this article didn’t mention was the loose policies surrounding peer review – for example, authors being allowed to choose who they want to peer review their studies. Stricter checks and balances in peer review need to be more firmly established to catch fraudulent studies before they even get published.
No comments:
Post a Comment