2018-06-26 09:44:45

Retraction Watch, a U.S.-based non-profit that records retractions of academic papers, has accused a Chinese anti-plagiarism website of copying its database.

Ivan Oransky, co-founder of Retraction Watch and a professor at New York University, told Sixth Tone that his team established their database at the end of 2016 as a reference for scholars. It includes papers retracted from journals all over the world for reasons such as plagiarism, doctored images, and faked peer reviews. Earlier this month, he noticed that a Chinese company called iPlagiarism maintained a database that was suspiciously similar.

According to Oransky, both databases are about the same size and share a similar web address — Retraction Watch operates retractiondatabase.org, whereas iPlagiarism’s address is retractiondatabase.com. “I find it very difficult to believe that they compiled these data on their own,” Oransky said in an email to Sixth Tone.

Retraction Watch first mentioned its database in a blog post on December 30, 2016. The database includes retracted papers from 3,944 journals and has more than 17,000 entries.

In a press release sent to Sixth Tone earlier this month, iPlagiarism said it published its database at the end of 2016, too. It includes papers from 3,702 journals, and its number of entries fluctuates between roughly 19,000 and 30,000.

Tan Xinjie, owner and manager of iPlagiarism’s mother company, Iplagiarism (Beijing) Science and Technology Co. Ltd., told Sixth Tone on Monday that part of their database is built on resources from Retraction Watch. “According to a notice on their website (for some reason they pulled the notice off recently), their data is free for non-commercial use,” he said through e-mail.

Established in March 2016, Iplagiarism provides paid plagiarism checks for both texts and images using big data and artificial intelligence. Their retractions database is a free service.

Oransky said he had been in contact with Tan earlier in June and that Tan had admitted to plagiarizing Retraction Watch’s data. By June 14, all data had been deleted from iPlagiarism’s English version, but not from its Chinese version.

When asked by Sixth Tone whether his company had plagiarized Retraction Watch, Tan said that his company’s database uses data from a variety of websites. They had not acknowledged Retraction Watch because iPlagiarism has “too many sources,” he said. Instead, their website carries a notice saying copyright of the original data belongs to their owners, who should contact the company in case they disagree with their usage.

Oransky said Retraction Watch is happy to cooperate with other parties and make its data available, and that the notice on non-commercial use can still be found on the website. “The impression [Tan is] giving is that we have changed our policy, that’s false,” he said, “as is so much of what he says.” The NGO is not taking any legal action, as they “don’t have the resources,” Oransky said. “It’s just very disappointing.”

While awareness of the importance of scientific integrity is rising around the world, Oransky said, misconduct in academic papers is still prevalent in some countries, including China, where the issue is compounded by a lack of record-keeping. Last year, 107 papers written mostly by Chinese authors were retracted from Tumor Biology because the medical journal’s publisher, Springer, discovered peer reviews had been faked.

On June 4, iPlagiarism published a report about their database that showed how retractions of Chinese papers were on the rise.

Oransky questioned the quality of iPlagiarism’s Chinese database because its size changes “by thousands of entries every day.” On June 16, NetEase Data News published a report based on iPlagiarism’s database, counting a total of 30,002 entries. As of Monday, however, the database had a total of 25,201 entries. Tan said the database is constantly being updated and called the information “very reliable.”

This article has been updated to include further comment from Ivan Oransky.

Editor: Kevin Schoenmakers.

(Image: E+/VCG)