Every scientist has received — and many have written — a negative review. Indeed, calling out unworthy manuscripts or shoddy experimental methods is a critical step in the peer review process. Nobody turns cartwheels on getting such feedback, but by and large, the community agrees there's a need for insightful criticism.
But there are bad reviews, and then there are bad reviews. You know the ones I mean: excessive demands for more experiments, blatant bias, needlessly adversarial comments. While constructive reviews help move science forward, cutthroat reviews can derail it.
On some level, everybody knows that. So why do we hear about reviews like this again and again? There's no trend data on this, of course, but anecdotally, indications abound that these slash-and-burn reviews are more common during tight funding times. In the era of the lowest paylines on record, then, let's consider what can be done to make the review process more about good science, and less about gamesmanship and fear.
For starters, we have to think about how scientists learn to write reviews: They don't. You'd like to think there would be a formal training procedure for young scientists to undertake such a tremendously important part of the scientific process, but what we're faced with instead is a combination mentorship/fly-by-the-seat-of-your-pants situation. Most scientists get their first reviews when they publish their first papers as grad students or postdocs, and then piece together how reviews should be written from those. As a result, take-no-prisoner reviews can turn into study guides, teaching the next generation of scientists that this is appropriate behavior in the peer-review process.
Veronique Kiermer, executive editor and head of researcher services at Nature Publishing Group, encourages scientists to include their trainees in the review process early on. Getting students to help draft parts of a review and go through them is one way for more senior scientists to mentor their trainees. (Those students should then be named as contributors to the review, Kiermer notes.)
Paul Flicek, a senior scientist at the European Bioinformatics Institute, tells his students to write reviews as a report to the journal editor, not as direct feedback to the author. Writing with the author as your intended audience tends to make reviews "more adversarial" in tone, he says. Another challenge he sees among reviewers is with those scientists who feel that the burden is on them to prevent any kind of fraud or other bad science from being published. "You get the feeling that they have appointed themselves guardian of the content," Flicek says. "That's where crazy requests for additional experiments and things like that sometimes come from."
Flicek, for one, is an advocate for named peer review. If people had to sign their names to reviews, he contends it "would make reviewing better almost instantaneously." Of course, that opens another can of worms entirely, but the anonymity of peer review certainly increases the opportunity to submit a terrible review.
As neither a scientist nor a journal editor, I'd like to offer some totally unsolicited advice to both parties.
Scientists: Write each review as if you planned to sign it, and don't direct your comments to the author. When you've finished with the review, read it one last time as if it were a review of a paper you've submitted, and use that to determine whether the comments are constructive or not.
Editors: You have enviable access to scientists — start using that to offer training on how to review, not just how to get published in your journal. And include in your reviewer resources a template of what an ideal review ought to include.
Meredith Salisbury is editor in chief of GenomeWeb. Feel free to disagree with her at [email protected] The views expressed in this column are not necessarily those of Genome Technology.