Poll: Research - to preprint, or not to preprint?

[Andy]

The broad question asked here is "Do you think that research should be published with, or without, a preprint stage?".

We have recently seen the disappointment of the NIH intramural study, where the authors were determined not to preprint, and so we were unable to get prepublication sight of what was, or wasn't, contained in the paper.

A counter-argument might be that peer review is an opportunity, in part, for errors made by the authors to be spotted and corrected before publication, and therefore it is better for that to happen before the public see any detail at all.

A counter-counter-argument might be that preprint allows for a wider range of input that then might be able to improve the future publication.

There may well also be additional arguments for and against, add them in the comments below if you think of them.

 

[Arnie Pye]

Because of paywalls, the preprint version of a paper is often the only chance I have to actually read about new research.

 

[chillier]

Generally speaking journals as they currently exist are problematic. They force scientists to pay to submit their own papers - funded often from public money, to a journal that then hides them from the public and charges the scientists again to read back their own work. It is also peer reviewed by other scientists - so the journals don't do the work but reap the rewards, as well as controlling what output the public sees.

Additionally you can not always trust that your reviewers will be unbiased, they may have a vested interest against the arguments the authors make, or even use the opportunity to see your work before their own is published, to scoop you. There have also been instances of peer reviewers insisting that you cite their work in your paper (corruption).

I'm very pro preprint despite the flaws you mention because it's democratising. Perhaps there could be some system where preprints are reviewed and their comments attached to the preprints publicly. So people not in the know can look to them to get a sense of what works and what doesn't in the paper, and people who are in the know can make their own minds up about if the reviewers are making fair and constructive criticisms.

EDIT: Also they set up bad incentive structures. Bad scientists who publish sensational results will fair better as their research is more marketable. Good scientists who report frankly and honestly with less sensational results may have a harder time publishing, pushing them out of science. Publishing negative results is important but undesirable for a journal.

 

[EndME]

There should always be a preprint. There really isn't a reason to not submit a preprint (there might be the very odd occasion where a meaningful argument can be made that before making something public there isn't allowed to be a single error anywhere, however in most scenarios your peers will anyways be roughly aware of your work via seminars, conferences or by talking to them and fundamental flaws would often have shown up already). The preprint is just the first version made public.

Field-medalists and nobel-prize winners in physics upload their work to preprint servers, there's no reason why biomedical researchers shouldn't. You'll find mistakes in every published paper as well. The question is always id someone finds these errors and if they influence the results at all. Submitting a preprint only makes it more likely for errors to be corrected (and preprints get updated and corrected as well, in fact it's much easier to correct a preprint than it is to correct something that is published).

That doesn't mean there shouldn't be a peer-review phase followed by a publication, this is crucially important as well (even if the publishing landscape is highly problematic). There's several problems currently going on with journals, however preprints only take some of that power away from the publishers and make it accessible to the public who are funding all of this work.

I also feel that people that start submitting preprints quickly become convinced of doing so and subsequently even sometimes go on to only publishing in open access journals afterwards.

Many mathematicians describe the Arvix as the open source collection of all mathematical knowledge, accessible to everybody. A library in the 21 century for which nobody needs a library card. That alone is a beautiful thought.

Publishing negative results is important but undesirable for a journal.

To me that seems to be a bit more of a problem in the whole biomedical field, rather than just a problem with journals, but certainly something preprints can be used extremely well for and maybe even help change this structure.

Edited:
Another advantage of preprints is that once you’ve submitted your work, the results are then “yours”, i.e. they offer protection against intellectual theft, which does still happen occasionally (There have been some very few reports, where one can’t be certain if they are true, of people stealing your work “via preprints”. That’s because most of these servers are managed by a select group of people that see preprints before they are uploaded and can then still produce work of their own before the preprints are uploaded. However, this risk is extremely small, can easily be overcome via different methods and I don’t think these very few cases were really “true” to begin with).

I know some authors also engage in preprint “popularity hacking”. By submitting your preprint at a certain time you can ensure it will on top of certain lists that get sent out to those who subscribe to them, which in turn means much more people will read and cite your work.

Even though they are both "problems", both of these problems only carry higher risks if you don't submit a preprint.

 

[Creekside]

I would vote in favour of preprints, but I admit I am totally ignorant of what mandatory preprints mean to scientists. There might be some factors I'm unaware of that make preprints bad for science. I'd like to see a debate among people who have an actual stake in the issue.

 

[Peter T]

A counter-argument might be that peer review is an opportunity, in part, for errors made by the authors to be spotted and corrected before publication, and therefore it is better for that to happen before the public see any detail at all.

Like @Creekside I am currently in favour of preprints but open to be being convinced of the opposite if there seem to be good arguments against. For me what sways me currently in favour of automatic use of preprints is the problems with peer review where reviewers seem to very often miss quite obvious problems in papers.

 

[Sean]

Peer review is clearly seriously broken in more than a handful of areas of science, and no more so than in psychosomatics, which has clearly gone rogue on methodology.

The situation is completely unacceptable. Something has to change. Lot to be said for at least trying pre-print reviewing.