Online exposure ‘leads to higher research paper correction rate’ | News | Times Higher Education

"Discussing potential flaws in peer-reviewed research online can lead to a higher rate of corrections to the scientific record, according to a study. The research, published in the open access journal PeerJ, found that life science papers containing data that are subsequently questioned in blogs are retracted or corrected about seven times more often than those that have been discussed privately. Paul Brookes, author of the research and associate professor at the University of Rochester Medical Center, told Times Higher Education that if academics identify potential problems in a published paper there is a 'benefit in going public'. Dr Brookes started an anonymous website, science-fraud.org, in July 2012 to flag up potential problems with scientists’ work. But in January last year he was forced to remove all content from the site and identify himself after he was threatened with legal action. He compared the number of corrections and retractions relating to 274 papers that had been featured on his blog with those of 223 papers with similar concerns that he had not yet posted online. The issues with each of the papers had been reported to the respective journals. He found that journals retracted 16 and corrected 47 of the publicly discussed papers, but only retracted two and corrected five of the papers where problems had been communicated privately. Overall, publishers applied 'some type of corrective action' to 23 per cent of papers that had been featured on blogs, compared with 3.1 per cent of privately discussed papers. 'It is therefore concluded that online discussion enhances levels of corrective action in the scientific literature,' Dr Brookes says in the paper, published on 3 April. He told THE that for publishers the research 'probably paints a less than stellar picture of the industry and the way they deal with these problems'. In the paper, Dr Brookes cautions that the findings may not apply to scientific literature at large because of the study’s small sample size and the fact that it focused mainly on image data from the life sciences ..."