Two years ago, I uploaded a preprint to arXiv.org describing what I considered serious problems, including apparently irreproducible results, that I had uncovered when analyzing a set of research articles published by the NASA Jet Propulsion Laboratory (JPL) NEOWISE project. NEOWISE is the largest scientific analysis of asteroids ever conducted; the researchers on the project have so far published estimated sizes of more than 164,000 objects in the solar system, estimates they have claimed were all derived by applying a standard approach to raw observations from the WISE space telescope.
Last year, chemist Marcus Tius at the University of Hawaii saw a paper describing the synthesis of some organic compounds, and was “struck by the implausibility” of the reported structures. So he joined up with some colleagues to try to replicate the data.
While Tius and his team were trying to repeat the experiment, however, in December 2017 the journal — Organic Letters — retracted the paper. The journal, published by the American Chemical Society, noted that the authors had not been able to produce crystal structures that confirm they had synthesized those compounds in particular. So Tius and his colleagues knew they couldn’t replicate the findings — but carried on their experiment anyway:
When geophysicist Craig Jones realized a figure in one of his published papers contained an error, he was on the fence about what to do. It was a clear mistake, but he’d seen much larger mistakes go uncorrected by other authors. Unsure if it warranted a correction, Jones polled readers of his blog to see what they thought.
More than two thousand researchers have signed a petition to boycott a new Nature journal over the fact it will be available only by subscription.
The new journal — Nature Machine Intelligence, about machine learning — plans to charge readers for access, unlike most other journals in the field. The researchers who signed the petition have pledged not to submit their work to the new journal, and will decline to review or edit papers for it, as well.
Most journals published by Nature Publishing Group are available only by subscription — but that doesn’t work for the machine learning community, the signatories argue:
Decades ago, unbeknownst to each other, two chemists were independently working on a screening approach to identify new potential drugs. Both published papers about the technique around the same time. So now, when scientists write papers that cite the technique, who should get credit for discovering it?
Decades later, that question still hasn’t been answered — and the researchers continue to argue, this time over one’s decision not to cite the other’s work.
A researcher whose work on the use of nanomaterials has been heavily scrutinized on PubPeer — with one critic alleging a paper contained “obviously fabricated” images — has lost eight papers. [Editor’s note: See update below.]
The eight articles — seven from Biosensors and Bioelectronics and one from Analytica Chimica Acta, both published by Elsevier — all cite issues related to duplications, and conclude with some version of the following:
Recently, a rash of news outlets posted concerns that canned tuna and other products may contain potentially dangerous levels of zinc. They were all wrong.
News outlets such as The Daily Mail and The Sun reported findings from a recent study, which showed that canned foods such as tuna may contain 100 times the daily limit of zinc — raising concerns about how such huge doses of the mineral could be causing digestion problems. The last author of the study told Retraction Watch the paper is going to be retracted, because the authors made a fundamental error calculating the amount of zinc present in canned foods.
After the paper appeared in 2017, one critic lamented it contained “obviously fabricated” images, and asked the journal to retract it. Another suggested the presence of one image merited “an instant lifetime ban.”
The authors of a highly cited 2016 research letter on a way to improve the efficiency of solar panels have retracted their work following “concerns about the reproducibility.”
Given the potential importance of the data, it would be nice to know what exactly went wrong, and why. However, the retraction notice doesn’t provide many details, and doesn’t even specify if the authors did indeed fail to reproduce the data.
Here’s an unusual way to allege plagiarism: Do it in the reference list.
That’s what Brian Levine, a professor in the College of Information and Computer Sciences at the University of Massachusetts in Amherst, did when he came across a paper he wanted to cite but suspected of plagiarism. When Levine published his 2017 paper, he cited the paper in question as:
R.Rajan, “ Feasibility, Effectiveness, Performance and Potential Solutions on Distributed Content Sharing System [plagiarized],” Intl. J. Engineering and Computer Science, 5(1):15638–15649, Jan 2016 http://www.ijecs.in/ issue/v5- i1/30%20ijecs.pdf.