Sunday, November 21, 2010

The Open Science unjournal

In this post I do something a little different than normal for this blog: I introduce a scientific unjournal called Open Science.  It doesn't exist, but it should.
 
What is an unjournal?
An unjournal is to journals what an unconference is to conferences.  To define what an unjournal is, take the first 2 sentences in the wikipedia entry on unconferences and substitute a few words:  "An unjournal is a facilitated, participant-driven journal centered on a theme or purpose. The term "unjournal" has been applied, or self-applied, to a wide range of publications that try to avoid one or more aspects of a conventional publishing, such as loss of copyright, high fees, [list your favorite pet-publishing-peeve here]."

Here are some frequently asked questions (FAQs) about Open Science:

How do I publish in Open Science?
The usual way is to deposit your manuscript on openscienceunjournal.org.  Once you have completed the submission process the paper is given a time and date stamp and the paper is published and open for review (see next question).  While there is a recommended template available for the paper, there is no fixed formatIt is possible to upload all your raw data or link to the data if it is hosted elsewhere.  Experience has shown that this tends to increase the scores (see below) of your papers significantly.

You can also choose to publish the paper on any of the unjournal sites whose contents are linked to openscienceunjournal.org.  These sites are often run by established publishers and offer more user-friendly interfaces, but may require a fee and may ask you to give up your copyright (though the content is open access by definition).

Is Open Science peer reviewed?
For a publication in an unjournal such as Open Science the question should be: is my article in Open Science peer reviewed?  That is in large part up to you.  The paper is open for online, non-anonymous, and completely transparent review and you have 2 months in which you can change the content of the article in response to the comments.  After the 2 month period you cannot change the content, but you can of course respond to new comments online.  For very serious criticisms you may want published a new Open Science article to respond.

It is up to you to solicit reviews, though any published author of a peer-reviewed paper (defined below) can review your paper during the 2 months review process.  Based on our experience only well-written and well-presented articles on scientifically interesting questions get reviewed.

An Open Science article is neither rejected nor accepted at the end of the review process.  Instead it receives an initial score (see next FAQ).  Obviously, any paper that does not generate a single review (or is reviewed but gets an initial score of 0) is not considered peer reviewed.

What is the impact factor of Open Science?
For a publication in an unjournal such as Open Science the question should be: what is the impact factor of my article in Open Science?  During the 2 months review process each reviewer gives the paper a score between 5 (good) and 0 (bad).  This score can be adjusted by the reviewers based on the changes you make within the first 2 months, after that it is fixed. The initial score for your paper is the average of all reviewers final scores.  

If the paper is cited it receives an additional score (called the current score).  The score is determined by the number of citations, and if the citing paper is published in Open Science, the score is weighted by the initial and current score of that paper.  The authors of that paper can also choose to indicate how important your paper was to theirs.  If high scoring papers cite your paper in a positive manner, the current score of your paper increases. Self-citations are not included.  

Why should I review for Open Science?
The work you put into reviewing is now documented for all to see.  Have you contributed greatly to science by identifying Open Science papers with high current scores?  Do the reviews you submit carry more weight with the author and other reviewers as a result?  Some sites now list Open Science reviewers with particularly high impact as a kind of editorial board for the journal.

How should I cite an Open Science paper?
One suggestion is: Author(s), Title, Open Science, date of submission, initial/current score.  If you publish in Open Science using the suggested template, the current score is updated automatically. 

Why should I publish in Open Science?
There are many reasons:
(1) You retain the copyright and anyone can see the paper.
(2) Your paper is accessible upon submission. (Don't rush to publish though: you only have 2 months to get a good initial score).
(3) The impact of your paper is evident in the citation, but disconnected from the conventional impact factor of the journal you managed to get it in to.  The initial score of your paper can help the paper off to a good start, but your truly important papers will ultimately be identified by its current score.
(4) You choose the publishing format you like.  What's your pleasure? machine readable? interactive figures? link to raw data?
(5) Your paper is a living document: comments or questions continue to roll in on important papers and you can update links to your papers (related articles, a new data format) as you see fit.
(6) If you write a good paper, you will get more reviews (i.e. more suggestions and input) but the rantings of a single idiot reviewer will not prevent publication.  Isn't this the place to publish daring and ground-breaking work?  

So what brought this on?
The blogosphere: Egon Willighagen's latest post got me thinking about this particular idea, but the general problems it is trying to address was brought to my attention by many other blogs such as Michael Nielsen's The Future of Science and Is Scientific Publishing About to be Disrupted? posts; most posts by Peter Murray-Rust; Henry Rzepa's long fight to include interactive figures in conventional journals; Mat Todd's excitement for an unconference and the discussion it generated at Derek Lowe's blog.  Why can't we have this in a journal?

Is the Open Science unjournal a good idea?
The blogosphere will decide: no comments on, and no re-blogging of, this post will mean this idea dies a quiet death (by receiving an initial and current score of 0).  But if you get enough smart people fired up about an important idea that can be solved by IT, good things can happen.

12 comments:

Marcus said...

Brilliant idea bravo! As a librarian I strongly support moving in this direction.

Marcus D. Hanwell said...

Great post Jan - I am certainly very interested, and think that this is the direction scientific communication should be moving in. Coming from physics, and always having arxiv I think something like this is needed.

Jan said...

Very happy to hear you both support the idea. Thanks for letting me know.

Matt Hodgkinson said...

It's been done, it is called Philica. It didn't work, at least I wasn't impressed with the result: http://journalology.blogspot.com/2007/03/tags-track-growth-in-open-access-and.html

Jan said...

Thanks very much for the link. I hadn't heard of Philica, and I agree it is not a success: 206 entries since 2006, and much of it garbage.

While it didn't work, I'm not sure this means that the general concept is flawed.

There is definitely some things I would do different, and that may have contributed to Philica's failure, such as the inability to change the manuscript in response to reviewers, and the lack of any "impact factor" in the citation, as I propose.

The main issue for any new journal is credibility, which in this case is exasperated further by the fact that pseudoscience will sneak in without editorial control. This is why I proposed that scores be part of the citation.

On a slightly positive note, take this Philica article, which would otherwise probably have ended up in a low-impact nanoscience journal, where I doubt it would have been viewed 13,305 times.

Jan said...

PS: "pseudoscience will sneak in without editorial control."

This apparently also happens with editorial control (and respectable impact factor): see here and here (source).

Henry Rzepa said...

It is intriguing to ask how the science reported in 10.1126/science.1188002 might have fared in an unjournal. In fact, since July 2010 when the article cited first appeared, it has been discussed on a variety of fora (for pointers, see here). Another (open) experiment on this very topic can be found here, with an outcome that we hope to present in one way or another at the ACS Anaheim meeting in March 2011.

Part of the problem of course is that some reviews of a scientific discovery are largely opinion (angels dancing on the points of needles and that sort of thing) and some are based on (and properly cite) facts. Others argue that the mere existence of debate implies the science has made an impact and is therefore important.

The story that is recounted in the links above still has a little way to roll, and it will be interesting to see what the verdict is once the passage of time has had its effect.

Jan said...

Yes, good example. I wonder if the paper would have generated as much discussion if it hadn't appeared in Science, but then again it might have generated more direct discussion with the authors in an unjournal.

It's an example of a publication that will generate more citations because of ambiguous data, than it would have if the data was rock solid. That's one reason I suggested a way for citing authors to somehow indicate why they cite the paper, e.g. do they agree of disagree with the paper they are citing?

In any event, the ensuing discussion will certainly be more productive in an open environment with sharing of data, such as the Quixote collaborative project.

Jedrzej said...

Great intitative!

I've reblogged your post - with sue acknowledgement - on my blog (polariton.wordpress.com). I hope that is acceptable - if not, let me know and I'll take it down.

Jan said...

That's no problem at all. In fact I am honored.

Matt Hodgkinson said...

I don't think those PLoS ONE studies that received criticism in the blogosphere were "pseudoscience". We see true pseudoscience and crackpottery submitted and we weed it out.

Studies to examine potential mechanisms of acupuncture are definitely science; whether there were issues with the methods or the reporting in those particular studies is another matter, and one PLoS would be very happy to receive further comments on: http://www.plosone.org/article/comments/info%3Adoi%2F10.1371%2Fjournal.pone.0011907; http://www.plosone.org/article/comments/info%3Adoi%2F10.1371%2Fjournal.pone.0012619. Refusing to allow the publication of such studies per se on the grounds that 'acupuncture is quackery' would be dogmatic, not scientific.

As for the modelling of wind setdown, that's a scientific study. It was quite a nice case study for the authors' model. People might not like the motivation for the study or what they read into the results, but that's not really a scientific issue. The study annoyed both strident atheists and evangelical Christians in equal measure...

Jan said...

Let me start by saying that I have nothing against PLoS ONE. On the contrary, you may receive papers from me soon.

I also agree that the mentioned PLoS ONE papers are less pseudo-scientific (i.e. sounds less crazy) than many of the papers in Philica that you mention in your blog post.

But it is the stigma associated with publishing such papers (i.e. the "I won't publish in a journal that publishes papers like this!") that unjournals have to overcome.

So it is nice to see that this doesn't seem to hurt PLoS ONE very much, if at all.