Friday, October 11, 2013

Real Peer Review

Over the past fifty years or so I have become less and less enamored with peer review. The problem? First is the anonymity. Who is this person? Why do they think they know the issue better? They may, but then I would like to know. Better yet are reviews that flatly state that all of this has been done before. Oh, but where? They never say. What of peer review of peer review. And the list goes on.

In a recent set of articles in Science this is all discussed in light of Open Access publishing. They state:

"Peer review is sick and collapsing under its own weight," he contends. The biggest problem, he says, is the anonymity granted to reviewers, who are often competing fiercely for priority with authors they are reviewing. "What would be their reason to do it quickly?" Tracz asks. "Why would they not steal" ideas or data? 

Anonymous review, Tracz notes, is the primary reason why months pass between submission and publication of findings. "Delayed publishing is criminal; it's nonsensical," he says. "It's an artifact from an irrational, almost religious belief" in the peer-review system. 

As an antidote, the heretic in January launched a new venture that has dispensed altogether with anonymous peer review: F1000Research, an online outlet for immediate scholarly publishing. "As soon as we receive a paper, we publish it," after a cursory quality check. Peer review happens after publication, and in the light of day. F1000Research selects referees, who post their names and affiliations alongside their critiques. Papers become like wikis, with reviewers and authors posting comments and revisions as the need arises. 

I have advocated this before. Peer review is a non-anonymous process post public release. What does this accomplish? The following:

1. You get to judge the author and the reviewer side by side. 

2. Reviewers with an axe to grind are exposed.

3. Anyone can review the piece subject to certain rules. Anonymity is prohibited. You must list your name and contact information. Your review must contain details and data and backup.

4. The Editor still plays a role but a role of orchestrating the reviews. Some reviews will be just junk, as usual there are always people with "stuff" to say. Let then say it elsewhere.

5. Self policing is essential. The readers should police the comments so as to drive out the bad ones, again policing must never by anonymous. If you decide to vote out a comment then you do so in the light of day. There has been the argument that Grad students could never put out bad comments on say a Department Head. But others could and should. Perhaps the truth will out.

6. It then begs the question of; what is the paper? Is it the original posting or the amalgamation of it and all the comments? If so then who is the author? This is a challenge, yet one we face all the time.

The article concludes:

Tracz acknowledges that in reshaping peer review, he's taking on a sacred cow. "There will be some growing pains," he says. But his maverick ideas tend to become mainstream over time. "At the beginning of open access," one colleague says, Tracz "was ridiculed by other [publishers]." No one ridicules open access now. 

"He's not radical," Eisen insists, "just sensible. Sensible doesn't [usually] happen in scientific publishing." The coming years will see whether open peer review is sensible—or too radical for most researchers to stomach. 

The comments above are suggestions based on thinking about this for a while. The  driver is the change in the distribution channel for scientific data. Namely the Internet changes everything. One can put things out there and see what happens. I have placed Drafts on line and have seen tens of thousands of downloads on several ones. Who, why, what did they think of it? That process is not working yet as one would like but the spread of the information does work.

A change in peer review is essential. Today it protects the interests of a few. It should be useful by expanding the interests of the many.