Small gap found in error reporting at JACS

More than 97% of the studies that have been retracted or corrected by the Journal of the American Chemical Society (JACS) in the last 2 decades are appropriately labeled, a new study finds. Still, almost 20 retracted papers were mislabeled as corrections, the analysis reveals.

The study is due to be presented in September at the 27th International Conference on Science, Technology and Innovation Indicators in Leiden, the Netherlands. The analysis is part of the European Commission’s Nano bubbles project, which aims to explore how, when, and why science fails to correct itself. JACS is published by the American Chemical Society, which also publishes C&EN.

Study author Frédérique Bordignon, who studies research integrity and scientometrics at École des Ponts ParisTech in France, examined 1,083 editorial notices issued by JACS between 2000 and 2023.

According to a 2010 study, JACS had to correct 1.4% of the more than 17,000 papers it published between 2000 and 2005. The new study found that 97.4% of the papers associated with a retraction or correction are correctly displayed on the JACS website.

“I don’t think we can be happy with these numbers,” Bordignon tells C&EN. “We need 100% because we need to be 100% sure of the papers, the content, the data, the images we are using for further research.”

In her research, Bordignon found 18 corrections that should in fact have been labeled as retractions on the JACS website. What’s more, out of these 18, 17 are labeled correctly as retractions on PubMed and 10 on Web of Science. That suggests a lack of consistency across different databases and repositories.

“It’s a mystery to me,” Bordignon says. “I was expecting the publisher’s source to be the best one so that we can rely on and be sure that all the other sources need to refer to that one source to be accurate, but that is not the case. For instance, PubMed is sometimes more accurate than the publisher itself.” Erick Carreira, the editor-in-chief of JACS, didn’t respond to requests for comment.

Bordignon says the inconsistency is a problem because researchers may come across the same paper from different sources and not be clear on whether it’s retracted or not. She attributes the inconsistencies between publisher sites and indexing repositories to a lack of time, effort, and resources being dedicated to make sure the correct information is displayed on all sources.

Todd Carpenter, executive director of the Maryland-based National Information Standards Organization, agrees that editorial notices need to be communicated consistently across different publishers. Carpenter’s organization is working on recommendations about how to communicate retractions, corrigenda, errata, withdrawals, expressions of concern, and other types of notices.