|Mark Ware: why is peer review an enduring factor?|
Peer review is not broken in his mind. It is overwhelmingly supported by researchers, but that doesn't mean it can't be improved. Publishers need to take new developments into account.
Peer review is one part of the publishing cycle as well as the broader research cycle. It is important. Benefits include improving the quality of published articles (and this comes from the research community). It provides filters, seal of approval and a reward system that works for researchers.
However, no human activity is perfect so what are the limitations? They include validity, effectiveness, efficiency, the burden on researchers and fairness. In a world where there's a drive for transparency, we have to take these criticisms seriously. Peer review is ripe for improvement in many areas.
So who's who in peer review?
- Authors (co-authors, contributors) - the definition of authorship has become more formalised in recent years.
- Editors and editorial boards - editor role is crucial, misconception that it is the reviewers that make the decision. This ignores the fact that peer review is a process. Editors use their judgment. The best examples are a constructive discussion with the aim of making the paper the best it can be.
- Publishers and editorial staff - these roles are often overlooked by those that claim reviewers do all the work. Editorial independence is a reason why we might want this. Peer review process diagram: the danger is that we think this is the system, but actuallyit is one small part of a publishing process.
- Readers (post-publication review).
|Peer review flow chart: just one part of process|
Retractions are booming:
In 1977 there was c.2 retractions per 100k publications.
In 2010 this had risen to 50 per 100k publications.
What's driving new approaches?
What are the problems we're trying to solve?
Which problems are we trying to mediate?
What are the opportunities to improve? Fairness and bias, delays, inefficiency, reproducibility, research data and information overload all figure.
Pre-publication innovations include:
- 'Soundness not significance'
- Cascade review
- Portable review
- Open and interactive
- Registered reports.
Post-publication innovation includes:
- Comments and ratings
- Metrics and altmetrics
- Article evaluation and tiering systems (e.g. Frontiers)
- Overlay journals.