Fewer cases of faked data this century, but significantly more cases of plagiarism/duplication
The retraction rate for genetics studies is around eight times higher than that for other life science disciplines, suggests the first research of its kind, published online in the Journal of Medical Genetics.
But the reasons for retraction⎼caused by loss of confidence in the study findings/conduct or in the publication process⎼have changed in recent years.
Plagiarism (lifting content produced by someone else and passing it off as your own) and duplication (the same content repeated in several publications) have become significantly more common prompts for retraction than faked (falsified/fabricated) data, the analysis shows.
Genetics covers a broad range of disciplines, encompassing both medical and non-medical specialties, such as anthropology, biology, and botany. It has a long history, but has experienced unprecedented development in this century, say the researchers.
They therefore wanted to find out how many of the genetics studies published between 1970 and 2018 might have been retracted, and the reasons why.
They retrieved information on 1582 retracted genetics articles from RetractionWatch, an open access database that has been tracking corrections and retractions in scientific and biomedical journals since 2010.
Of these, 1443 were from seven countries selected for closer study: China, Germany, India, Japan, South Korea, the UK and the USA.
The researchers also searched for any articles that had been subject to civil or criminal proceedings, or investigation by the institution/company/publisher, other third party, or the US Office for Research Integrity.
More than four out of 10 retracted articles (44%) were medical; the rest were non-medical (56%).
One in three (33%) of all retracted articles involved research misconduct, such as falsified/fabricated data and plagiarism. Nearly one in four concerned duplicated content. Nearly four out of 10 (37%) were investigated.
A higher proportion of retracted non-medical genetics research contained fabricated/falsified data (28%) than medical genetics research (18.5%). But retracted medical genetics research articles were significantly more likely to be investigated: 45% vs 31%.
Most retracted articles from among the seven countries scrutinised were authored by researchers from the US (526; mainly non-medical genetics research) and China (509; mainly medical genetics research).
The lowest numbers of retractions were for articles authored by researchers from South Korea (64) and the UK (69). Certain authors were serial offenders.
Analysis of time trends showed that data fabrication/falsification was a significantly less frequent reason for retraction in 2011-18 than it was in 1970-2000.
On the other hand, plagiarism and duplication were significantly more common reasons for retraction in 2006-18 and 2001-18, respectively, than in 1970-2000.
This might be because image editing software now makes it easier for researchers to duplicate content, but it also makes it easier to detect, suggest the study authors.
The average time it took to retract medical genetics articles was considerably shorter than it was for non-medical genetics articles: around 5 vs 6.5 years.
But the overall time to retraction for research involving fabricated/falsified data and plagiarism shortened significantly in 2006-18 and 2001-18, respectively, compared with 1970-2000, although articles involving faked data took up to three times as long to retract as those involving plagiarism.
Between 1996 and 2017 some 975,000 genetics articles were published, of which 1476 were listed in the RetractionWatch database, adding up to a retraction rate of around 0.15%, calculate the authors.
This is “almost four times higher than the current rate found on the [RetractionWatch database] for all disciplines (0.04%) and almost eight times higher than that found for PubMed articles (0.02%),” they write. “Therefore we should assume that genetics is a discipline with a high retraction rate,” they conclude.