Weekly digest: what’s happening in open science?

Amy Williams

Featuring the push to shine a light on peer review reports, worries about preprints in the press, and research on trial reporting levels in the top society journals. 

Is it time for peer reviews to come out of the dark? via Nature

Making peer review reports public is not a new idea by any means, and several journals such as F1000Research, Royal Society Open Science and Nature Communications already publish peer review reports alongside the article they critique. Discussions ahead of Peer Review Week have urged other journals to follow suit. In an open letter penned by ASAPbio’s Jessica Polka, a collaboration of journals with open peer review urged others to follow suit. There are many benefits to open peer review; it not only adds transparency but also offers important expert insight into the subject matter, and allows peer reviewers to gain credit for the hours they put into reviewing articles without financial compensation. However, some researchers still harbour reservations and fear that open peer review may lead to reviewers being less honest when reviewing papers authored by senior academics out of fear it could damage their career prospects.

Do medical journals do enough to ensure clinical trials are reported? via Trials

By registering a trial before its start date and including the hypothesis being tested, endpoints and planned analyses, the trial results can be protected against biases, unconscious or otherwise. This paper explores the rates of adherence to the clinical trial registration policy of the International Committee of Medical Journal Editors of the 50 most recently published clinical trials in the ten highest impact medical journals run by medical societies in the USA. The authors found that 10% of the trials examined were unregistered, and almost a quarter of those that were had been registered retrospectively. Trials funded wholly or in part by industry were more likely to have been registered prospectively compared with those with non-commercial funding. They also found that unregistered trials were more likely to report favourable findings (89% positive for unregistered versus 64% for registered trials). This study emphasizes the importance of trial registration in reducing bias, and calls on journals to work harder to enforce pre-registration policies.

Can the press be trusted with preprints? via The Wire

Everyone is familiar with the flaws of medical journalism – reports one day declare that something is good for you; the next, it causes cancer. While misleading, these stories are often ultimately based upon peer-reviewed papers, even if the outcomes had been exaggerated or the study was designed to look at something else. This article asks a simple question: with such misrepresentation on the loose, even for peer-reviewed articles, are preprints a good idea? Taking the example of a preprinted paper that made it into the mainstream media, only to be proven false by an academic in the field, this demonstrates the dangers of preprints for non-specialists. Concerns about preprints are particularly pertinent to health, where news coverage can exert major influences on behaviour and even legislation. Defenders of preprints argue that there are plenty of examples where bad science has passed peer review, and that preprinting papers may actually prompt journalists to view the results of academic papers with a little more scepticism.

Image courtesy of John Harvey