Tuesday, 12 June 2018

Separation of review powers into feedback and importance assessment could radically improve peer review

After some blog posts about grassroots journals, it looks as if no one else will pick up the idea and I have started creating a first grassroots journal.

(It is interesting how often fear of being scooped is mentioned as reason against Open Science. Typically good ideas are not recognised before they are presented in detail and even then it takes time. At least that is my impression with the small paradigm changes I was responsible for: surrogate clouds and adaptive parameterisations.)


"It was the pits," said [economist Brian Arthur]. "Nobody there believes in increasing returns."

Susan Arthur had seen her husband returning from the academic wars before. "Well," she said, trying to find something comforting to say, "I guess it wouldn't be a revolution, would it, if everybody believed in it at the start?"


The first grassroots journal is naturally on homogenisation of climate observations. That was a good way for me to check whether the idea would work in practise. I think it does. And a concrete example is a good way for everyone to see how such a journal would work.

Using a WordPress blog works, but I am learning how to code a WordPress site to make it more user friendly and to make it easy for everyone to start such a grassroots journal, just as easy as starting a blog. (Which is really easy. Hint.)

With that it also became time to spread the idea and I have written a short guest post for the blog of the OpenUp project, an EU project on Open Publishing. After all these years of blogging that was my first guest post.

I was supposed to write about how wonderful openness is. So I wrote about how wonderful the right mix of privacy and openness are. Scientists are natural contrarians.

The main message, however, was already presented in my last blog post that if we separate the two roles of peer review — 1) feedback for the authors and 2) advising the journal whether the article is important enough for them — we will get a much more healthy quality assurance system.

In the feedback round I see no real reason to publish the content, all the (little) mistakes that were corrected. However, as it is just feedback, a friendly helping role, it would be easy publish the name of the reviewer.

In the assessment round the review itself is very interesting for others as well and is best published. Because this is judging a colleague and anonymous makes it easier to be honest. I would give the choice whether to be named to the reviewer.

Enjoy: Separation of review powers into feedback and importance assessment could radically improve peer review.

2 comments:

aduesterhus said...

Sorry, but I cannot agree with you on that point. As you know I have written my PhD about publication processes and have covered therein also this topic. So basically (also published as Quadt et al 2012) you can make the distinction between quality assurance (Copernicus calls it discussion paper or Open Peer Review, others call it preprint server) and peer review (so one or more independent reviewers are selected by an editor, who makes a decision on publication based on their reviews). Nevertheless, to take the QA step completely out of the PR step is a dangerous game. What happens when the QA in Open Review fails (as it is hardly ever really seen serious nowadays) and the peer reviewers are just allowed to decide on the fitting to the journal? No, what you propose here is basically guaranteed to be seen as a grey literature journal and so does not qualify as a scientific acceptable journal.

Victor Venema said...

Did you also write about publishing scientific articles? I only knew you were working on publishing data and how to help those reviewers. We should talk.

I would not even call the open discussion papers of Copernicus "quality assured". They are just manuscripts. An editor may have had a quick look at it, but will normally not have the expertise to see any problems. Quadt et al 2012 would be bending the English language somewhat.

Thus I would argue that what I write about above is about the peer review and a proposal to split that up in two phases. The split in manuscript publishing and review is already there with preprint servers (and Copernicus).

There would be two rounds of review. So it is not clear to me why you think there would be no Quality Assurance in this scheme (the above post is just the introduction to the real post).

If someone has no expertise it is naturally a bad idea to read papers after the first (feedback) round (or manuscripts) before they passed the second quality (assessment) round.

Quadt, F., Düsterhus, A., Höck, H., Lautenschlager, M., Hense, A. V., Hense, A. N. & Dames, M. (2012) Atarrabi – A Workflow System for the Publication of Environmental Data. Data Science Journal 11, 89-109.

My apologies I released this comment so late, there was somehow no email warning that it had arrived.