This session by Ken Carslaw from the School of Earth and Environment for University of Leeds. He is using Atmospheric Chemistry and Physics journal as example.

Atmospheric Chemistry and Physics (ACP) was founded in 2001, run by the European Geosciences Union (EGU). It now has the highest Impact Factor of the 4 Atmospheric Science journals listed by ISI.

Success (they believe) due to:

• Open access
• Collaborative peer review and commenting (the most innovative feature in the journal)
• Speed of publication (two stage publication, with submitted papers available immediately they are submitted)
• Flexibility (special issues, new article categories, etc.)

A lot being covered in the talk today is available in Poshchl, Learned Publishing, 17, 105-113, 2004

The process started with some basic issues/realisations:

Traditional peer review not an efficient means of quality assurance

• Limited capacity/competence of editors and referees
  • few editors for large subject areas – limited knowledge of scientific details and specialist referees
  • work overload, conflicts of interest and lttle reqard for referees
• Retardation and loss of information in Cloased Peer Reivew
  • The right person doing peer review on a paper can make a real contribution – but within closed peer review it may not go to that person (in fact it is unlikely)
• Spare and late commentaries in Traditional Discussion
  • cf Faraday Discussions – you circulate paper well in advance, then discussion ‘in the round’ at a meeting where paper is presented – get dialogue – they wanted to get that in the published paper environment
  • comment/article ratio has dropped significantly in the last 30 years

Large proportion of scientific publications carelessly prepared and faulty

• Fraud (rare)
  • selective omission, tuning and fabrication of results
• Carelessness (frequent)
  • superficial and irreproducible description of experiments and models
  • non-traceable arguments and conclusions, duplicate and split papers

By exposing papers on the web for open peer review, researchers are more careful, as aware that the work will be ‘public’ and reflect on them.

Conflicting needs of scientific publishing: rapid publication vs. thorough review and discussion

• Rapid publication – widely pursued
  • brief papers, rapid reviews, curtailed review and revision process
• Thorough review and open discussion – still the exception
  • required to identify scientific flaws and duplications
  • traditionally limited by availability of referees, review time and access to information

Came up with two stage publication with collaborative peer review:

Stage 1 – rapid publication of ‘discussion paper’ (D-paper) – passed by editors, full citeable, typeset and permanently archived

The paper is typeset at this stage, paginated etc.

Followed by public peer review and interactive discussion – to anyone registered – these discussions are also fully citable, and often are, as they contain important information

Stage 2 – review complete, final publication.

Questions about ‘Discussion Papers’

• Should D paper be paginated?
  • Yes – so it can be cited
• Should the archive D paper be a ‘journal’?
  • Yes, so it can be cited (not grey literature)
  • Not ISI listed – by lose citations, as many people cite the ‘D’ paper, and don’t bother to cite the final version
• Should it be reviewed or accepted ‘as is’?
  • A minimum of quality assurance/filtering
• If the paper is eventually not accepted, should the D-paper be removed?
  • No, Impracticlal Deterrence (don’t want a non-approved paper hanging around)

Now Ken showing the workflow as a diagram… Noting that there has only ever been one instance where they withdrew a comment because it was simply an unsubstantiated attack (I suspect that specific papers could attract particular type of comments – e.g. the original paper on MMR and autism etc.)

The rules for ACP are:

• Peer reviewers >=2 – can be anonymous or attributed
• Public commentators – must be registered and are attributed
• Comments are not review or solicited – should be substantial in nature (although aren’t always)
• couple more rules I didn’t get…

To see an example go to ACPD website (use Google, I’m not online as I write this!) and navigate to “Most Commented Paper”

The advantages are:

• All win situation for authors, referees and readers
• Discussion paper
  • Free speech and rapid publication
• Public peer review and interactive discussion (collaborative peer review)
  • Direct feedback and public recognition for high uality papers
  • Prevention of hidden obstruction
  • Documentation of critical comments, referee disagreement, controversial arguments, scientific flaws and complementary information
  • Deterrence of careless papers
  • Special issues more collaborative process

Some interesting stats:

• Get about 5 submissions per month, with rejection rates of these at around 10%
• Final papers rejections run at about 10% (making 20% in total)
• Submission-to-publication time is 3-6 months in totla

Impact factor has increased steadily since ACP was established.

EGU (the publisher), has now established 8 interactive OA journals, as well as 2 OA journals with traditional peer-review and 1 subscription journal.

A very interesting model. The journal works on an author pays model – which is a pay on submission (unlike BMC which is pay on publication)

A question – how many of the papers are available in a repository – as far as Ken knows, they aren’t generally (although of course could be happening, and how would they know?)