More Responsible Peer Review and Less “Fantasy” in Science
(October 11th, 2016) Science publishing is not experiencing the best of its time. Nobel laureate, Thomas Südhof analyses some of its problems and proposes potential solutions in his most recent essay.
Scientific knowledge needs to be properly communicated. This was true in past centuries, when both the number of scientists and readers was much smaller. It is vital today, in the Digital Age, when more than a million scientific papers are published per year and are accessed by more people than ever. Science publishing is an essential part of the research process and failing to do it responsibly translates into problems within and outside the scientific community.
Thomas C. Südhof, biochemist at Stanford University and Nobel laureate, has recently reflected on this issue and offers his “personal perspective” about “truth in science publishing” in an essay published in PLoS Biology. He argues that the current peer review system is where some of the problems emerge. One of them, he suspects, is a conflict of interest usually hidden by both journals and reviewers: the former biased towards research that could potentially increase their journal profits and the latter sometimes following a personal or professional agenda while reviewing a paper. Moreover, Südhof complains about journals and reviewers not paying a similarly high price when they fail to review a paper properly, e.g. authors are usually the only ones to blame for papers that need to be retracted. Finally, he believes that a limited journal competition does not help to improve the quality of the peer review system.
The essay also addresses an issue that is so widespread in scientific publishing that we don’t realise its seriousness: exaggeration or biased interpretation of results. Every year, the scientific community is scandalised by a few cases of fraud but it is not uncommon to find, each week, published papers with conclusions that do not accurately match the data presented. As Südhof suggests, publications ought to “report facts and not fantasy” and, if authors are failing to draw proper conclusions for their work, then journals and reviewers must demand it.
Südhof is also concerned about the reproducibility of current scientific research, an aspect not always treated as carefully as it should be. Examples of this are papers that, with only limited results, give misleading statements, or those that fail to perform proper experimental controls. There are also the cases of experiments that cannot be replicated due to the complex conditions, in which the original work was done or because the methods and materials are not accurately described.
The panorama feels overwhelming but can we actually do something to improve the situation? Südhof suggests the following ideas. First, higher accountability for journals, editors and reviewers is needed. Making reviews and reviewers public could be one action to do so. Second, we must improve the quality of evaluation for papers. Submitted works must meet characteristics not always assessed today. For instance, reviewers should care more about reproducibility and encourage the presentation of solid evidence, backing a statement rather than accepting exciting discoveries that are not actually strongly supported.
These are not completely new ideas and some authors, journals and reviewers already work following some of the standards suggested by Südhof. Nonetheless, the scientific community must make further efforts to change the situation the science publishing system is currently facing. Improving the public credibility of science should start “at home”, by responsibly presenting and evaluating research work.
Photo: Steve Fisch