The nationwide experiment will initially include around 100,000 volunteers.
The journal is sharpening its review of life science papers and giving authors additional space to document more detailed methods.
April 25, 2013|
NATIONAL CANCER INSTITUTE, DIANE A. REIDThe prestigious journal Nature is making an effort to publish fewer irreproducible papers, it announced in an editorial yesterday (April 24). The journal will get rid of space limitations on Methods sections and will commission statisticians to help review some papers, among other measures.
The announcement follows widespread discussion of how to solve the problem of irreproducibility. “The problems arise in laboratories, but journals such as this one compound them when they fail to exert sufficient scrutiny over the results that they publish, and when they do not publish enough information for other researchers to assess results properly,” the Nature editors wrote.
The journal is encouraging scientists to be more transparent about their data, statistics, and methods, including posting raw data online. A checklist for authors submitting life science papers specifies necessary information about sample sizes, statistical tests, and blinding, among other topics. It also asks researchers to disclose sources and profiles of biologically variable substances, such as cell lines and antibodies.
Additionally, the editorial places blame for irreproducibility on inadequate statistical training for biologists and says that there are insufficient incentives for labs to try to reproduce results, as it is difficult to fund or publish such studies.