Researchers pay money, and hope for prestige. | Illustration: Melk Thalmann

Despite digitalisation, traditional academic publishing houses are still in control of their field today. Couldn’t researchers organise themselves instead? Then they could set the rules and publish their findings at cost price. Horizons has posed these questions to both established publishers and open-access advocates. It wasn’t an easy task, as it turned out.

Many publishers did not want to engage in the debate. Others did not even respond to our request. We contacted a total of nine publishers: big, traditional publishers and smaller, young publishers, including six in Switzerland and three from the rest of Europe. Two who agreed at the outset then withdrew when things became concrete. In the end, just two publishers took the trouble to answer our questions. Two researchers who are involved in publishing joined us to complete the picture.

Why do preprints have to be published a second time in an ‘official’, specialist journal?

Preprints have been a trend in science for a long time now, and they received another boost during the Coronavirus pandemic, when speed was essential. Suddenly, almost every paper on the subject was deposited on a so-called preprint server before later appearing in ‘classical’, specialist publications. The publication process therefore became more transparent.

Matthias Barton is a professor of medicine at the University of Zurich and the senior editor of the British open-access journal Elife. He’s sceptical about all this, because he maintains that preprints are merely unreviewed summaries of scientific papers. Specialist journals, on the other hand, reject bad manuscripts. When it came to papers on COVID-19, preprint manuscripts were being quoted in the media without it being made clear that they had not yet been peer-reviewed.

Kerstin Mork is a Senior Communications Manager at Springer Nature Group in Berlin, and she too sees a problem here. “Since there is no peer review, there is no guarantee whether the content of a preprint will stand up to an objective assessment”. It’s the publishers who guarantee the peer review process and thereby provide a reference point for further scientific work.

Should we uphold the peer-review process, even though it has repeatedly proved to be inadequate?

Since December 2020, all papers submitted to Elife have to be available beforehand on the Internet as preprint manuscripts. In this way, says Barton, an initial peer review already takes place before publication. However, the actual peer review still has to proceed when the paper is submitted, as this prevents bad articles from getting published. If a manuscript is accepted by the editor, it goes to a member of the editorial board who has the necessary expertise in the subject, and they then get reviews from external experts. The papers then go back to the authors for reworking before finally getting published. This reworking process has a positive impact on the quality of the article.

Does it make sense for peer reviewers to watch out for cases of fraud?

Time and again, cases of fraud go undiscovered, and are blamed on the supposed inadequacies of the peer review system. The science journalist Ralf Neumann of Laborjournal in Germany feels that this accusation is unfair. He wrote the following in a commentary on it: “Reviewers ideally have the task of assessing the work of their peers critically but also as benevolently as possible. They should not have to suspect every author of potential fraud right from the outset”.

If there are complaints against published articles or manuscripts, they are investigated by the editorial board. “If necessary, other people and institutions will be consulted, including university authorities and experts in the field”, says Stefan Tochev, who is the Marketing and Communications Manager at MDPI, an open-access publisher founded in Basel in 1996. They only check when complaints are made. The situation at Springer Nature sounds similar.

Publishers do check texts for falsified data, says Barton. But they are not in a position to do so in terms of an article’s content, and have to commission researchers to do this. “The problem is that there are often only very few proven experts in a specialised field, such as certain subfields in medicine”, he says. “In a worst-case scenario, the wrong data in medicine can cost human lives”. But at present, he says, you can publish almost anything as long as you pay for its publication. This is why he is campaigning for an independent investigative body to act as a regulator, checking possible instances of misconduct and sanctioning those responsible if necessary – in the manner of the Office of Research Integrity in the USA.

Wouldn’t it be more honest if peer reviews were carried out after publication, so that no one could make any accusations of censorship or nepotism?

There are many stories in science and scholarship about someone helping a friend to get published, about competitors thwarted behind closed doors, or about greasing palms to get a publication accepted. If the quality of an article were determined after its publication, then it might be possible to put a stop to this.

Kerstin Mork from Springer Nature disagrees. Peer review before publication helps to provide an objective assessment of a scientific article. After a successful peer review and publication process, the finalised “version of record” can be a possible basis for further work. If the peer review were to take place after publication, there would be a danger that erroneous results might be incorporated into the work of others, and a multiplication effect would ensue.

Would peer review for books be a good idea?
“The Swiss National Science Foundation already requires peer reviews for books today, which makes sense”, says Daniel Hürlimann, a professor of business law at the University of St. Gallen and the co-founder of the Swiss open-access book publishing company Sui generis. In disciplines that didn’t have any peer review until recently, such as law, people are still going through a learning process.

Should books be published digitally as a matter of principle?

According to Mork, Springer Nature has been publishing all its books as e-books since 2005, but they can also be ordered in print form. Hürlimann believes that academic books can and should be published digitally. However, many readers still like to have a printed book. Publication in both forms is best, and the actual printing process is only a small item on the budget for publishing a book. Now that the Swiss National Science Foundation funds the open-access publication of books, the costs of proofreading, typesetting and even for a so-called enriched e-book can be covered (the last of these covers additional functions such as audio and video files, and links to external websites).

Couldn’t articles and books be adapted continually to keep up with the current state of knowledge?

This model is something we know from Wikipedia, whose authors regularly adapt their work to keep up with current advances in knowledge, and where you can track the changes that have been made to an article. Barton thinks that this model makes sense, but only if you can ensure quality control by involving recognised experts. The platform Uptodate.com is a good example of this – it’s an online medical reference work by and for doctors.

Hürlimann believes it is essential to science and scholarship for a cited source to be consistent. However, this does not exclude the possibility of dynamically adaptable forms of publishing. “You just have to ensure that a specific version of a source can be cited and retrieved”.

How much should publishing cost?

The representatives of MDPI and Springer Nature did not reply to this question. The specialist journal Nature demands some CHF 10,000 for publishing an article on open access. Barton regards such sums as ethically indefensible, especially if you consider how few people actually read some articles. Ideally, publishing should be free, he believes, though it would perhaps be better to say that it should cost as little as possible, but as much as necessary. “We need a model for scientific publishing that moves away from private companies with their business models and back to the researchers who want to pass on new knowledge that has been validated”.