Preprint archives: why and where?

Here I collate some information about pre-print archives: what are they, why would you use them, why wouldn’t you use them, and which ones are best for ecology and environmental science. Again, a disclaimer: I am not an expert on this, this is mainly for my information, but perhaps also for you as well. As with my post on open science resources, I would recommend you discuss this with your librarian, if you have access to one.

Preprint archives can include versions of documents a) prior to or on submission to a journal (‘preprints’), b) once accepted to a journal, i.e. peer-reviewed but before type-setting (‘postprints’), or c) ‘reprints’ of formatted, peer-reviewed articles should the licence allow. Here I focus on (a) because they address the main benefits of preprints, and specifically focus here on repositories for manuscripts (not other dissemination products). Preprint archives have been around since the 1990’s. Some fields have embraced it (e.g. Math) whereas others are slower, but there are an increasing number of archives, including general, discipline specific, and geographic regional, and an increasing use of archives in the last 5 years. For a more detailed review of preprints in science (as of 2018/2019) see here.

Some benefits of preprints are that they provide open, timely access to and dissemination of scientific papers, by being open access, and removing the months or years that papers can spend in the peer review and publication process. They are citable (often with a DOI) and time stamped, which mean that they can be referred to and cited in papers and grant applications. This can be useful, for example, if you have a set of inter-related papers and you need to cite them without the vagaries of the publishing timeline, or if you need to report ‘output’ from a grant before a deadline. For early career researchers that are yet to establish a ‘career worthy’ publishing record, it can make your research visible in time for job interviews or key grants, and can signal that you are one of the forward-looking Open Science converts. It can also be useful to publish ‘negative’ results or ‘failed’ experiments, that can have a difficult time being published elsewhere. Your articles can also get more feedback from a more diverse range of commenters, and you can develop relevant networks of key researchers while you are still working in the topic, likely leading to more dissemination of the final published paper.

The reasons not to publish a preprint are getting fewer. One of the main reasons would be because a journal you want to submit to precludes this – however, this list is getting smaller over time, and you might want to ask if you really want to publish there anyway? Particularly with many funders now requiring open access publication, we are often now obliged to provide an open access version of an article (typically the accepted, pre-formatted version) if the final publication is not open access. However, I can accept that sometimes you want to reach a specific audience, and there are definitely things that need to be considered (e.g. see below) if you want to publish in specific journals to achieve this.

Some fears might be that ‘competitors’ can see the work and may ‘scoop’ research. However, preprints are a good way to timestamp ideas, and claim them publicly. This is a much more open process, compared to what might already happen by such unscrupulous competitors in traditional (blinded) peer review. One of the other issues might be that the article can never be “double blind” for review, however it rarely is a true blind anyway. My personal rational opinion is that this extra review can only make science better. Some reviewers could benefit from the help (e.g. reviewer 1 who’s sum total of comment is “seems alright”). Whereas it might help authors predict and address potential responses from others (e.g. the classic “reviewer 2”). It would be interesting to see how much preprints even get commented on anyway, a quick flick through some papers that are more relevant to me suggests this is not so common.

One of the big ‘issues’ with preprint servers is that many of the articles are not peer-reviewed, and sometimes media attention can be directed to non-peer-reviewed research that has no clearly defined caveats or is simply wrong. But the peer-review process is rotten and flawed, and often media attention is given to articles that are wrong, regardless. There is also the potential for journals to be hesitant to publish something that already has media attention. However I wonder if this is a) more relevant in medicine and not so relevant to ecology and environmental science, and b) not so much of an issue for journals as they get paid for access to the final article (which is typically the one cited and therefore ought to be read to be cited), and the more publicity the better. Similarly, there could be concern that the quality of articles might decline – however this seems not to be the case, as scientists value reputation. I would think that most pre-prints in ecology and environmental science end up finding a home in journals eventually.

Ok, so you want to publish a preprint. How to choose where to do this?

Considerations on where to publish a preprint might include (in no particular order):

  • Acceptable by target journals – typically journals will accept preprint publication (prior to submission) but have different policies on locations that are acceptable, and further conditions such as embargo periods for releasing accepted versions (search on SHERPA/RoMEO or wiki for journal policies – useful also as it provides links to the specific place in the publisher policy, not always easy to find otherwise)
  • Links (e.g. direct submission) to target journals – not essential but sure make things easier.
  • Moderation or peer-review – how much feedback do you want?
  • Discipline specific or general – discipline specific might have more exposure to relevant communities.
  • Scope and article types accepted – just like journals, they have specifications for this.
  • Licence options – consider what specific licences (or the absence of them) might imply.
  • DOI – this is pretty much an essential for pre-prints.
  • Revisions & ability to link the final published version. This is nice, especially if there is a lot of value added in the peer-review process.
  • Searchable or in search engines. If you want it to be findable through anything other than a direct DOI link.
  • Hosts supplementary materials – or links to services that do.
  • Citation/download metrics – These can be useful in these days as ‘performance metrics’.
  • Longevity (of the database) – while this is perhaps not so important if you intend it to be temporary, prior to final publication, it can be a consideration in many cases, especially if it remains the only open access option.

Remember, this landscape is changing rapidly, so it is always advisable to check individual journal and archive policies directly.

What repositories exist and what are some journal policies?

There are many lists kept by others, including ASAPbio (updated Jan 2020, described here), Joseph McArthur, and fosteropenscience. These different lists have different criteria discussed. Some of the more relevant for us/me are the discipline specific arXiv (physics, mathematics, computer science, quantitative biology, statistics, engineering, and economics), bioRXiv (biology), EarthArXiv (earth & planetary science), and SocArXiv (social sciences), while general repositories include Figshare, OSF Preprints, and Zenodo. Note, this field is rapidly changing. For example, PeerJ, is no longer accepting preprints: the numerous other preprint servers they’ve decided to focus more on publishing instead. Another example is that EarthArXiv moved from OSF to the California Digital Library due to longevity/funding concerns. Note also that publication on one archive can also mean it is searchable via others e.g. OSF preprints collate preprints via SHARE.

I might expand on this later, but for now here are a few more details on some of these (remembering that these details might not be 100% correct when you read this):

bioRxiv is intended for unpublished preprints in the life sciences (and is now separate from medical/epidemiology preprints). It is intended for manuscripts at the time of submission to a journal, and can accept revisions up until acceptance to a journal. Articles are screened so ensure they are appropriate, not plagiarised, and conform to ethics (e.g. animal experiments). They must be relevant to life sciences, and classified further into New, Confirmatory or Contradictory results (i.e.include experimental, mathematical or computational work, but cannot be reviews, theses, dissertations, student projects, or simple protocols). Authors retain copyright and can choose the option (CC BY, CC BY-NC, CC BY-ND, CC BY-NC-ND, CC0, or no reuse) – although note that one condition is allowing text mining. Articles are allocated a DOI, and cannot be removed (but can be withdrawn), and it will automatically link final journal publications to the record. Direct transfer to journals include F1000, the Frontiers-in series, PeerJ, PLoS, PNAS, Proc Roy Soc B, and Science, among many others. They are indexed by Google Scholar and Crossref, but not Web of Science. Readers can comment (comments are also screened). Usage is tracked via altmetrics, and abstract/pdf views.

EarthArXiv’s scope is earth and planetary sciences (including environmental biology and ecology with an earth or planetary science perspective) It accepts preprints and postprints, including research articles, review papers, case studies, technical notes (e.g. new instrumentation or new analysis methods), confirmatory studies, ‘null’ results (i.e. results that do not support a hypothesis), dataset description papers, and software papers. Articles are screened, allocated a DOI, can be subsequently modified (but not deleted), and are indexed by Google Scholar. It does not host supplementary materials (but encourages and links to Github, Zenodo, figshare, etc to do so). Articles are not peer-reviewed, but do allow for public comment. I couldn’t easily find relevant information re licences, or direct transfers.

Of the general platforms, more information on OSF preprints can be found here. It is my current, under-researched impression that Zenodo and figshare are possibly more intended and useful for non-preprint resources (e.g. talks, posters, reports, supplementary information, data and code) than manuscripts intended for publication, but someone please correct me if I am wrong.

Preprint policies from journals differ by publisher. Wiley journals, such as Conservation Biology and Journal of Ecology, and Journal of Applied Ecology allow preprints (prior to submission) on Author’s Homepage, Institutional Repository, or other repository (arXiv, AgEcon, PhilPapers, PubMed Central, RePEc, SSRN), although these must acknowledge acceptance for publication, link to publisher version. Preprints can be updated with accepted versions (given an embargo period of 12-24 months) but not the final version. Note, other Wiley associated journals like People and Nature have differing policies. Springer-Nature also accepts pre-prints (even hosted a repository for a few years), and has no embargo period on post-prints on an authors’ personal website, although this is 12 months for institutional or other repositories. This is similar for Elsevier and Taylor and Francis. IOP journals such as Environmental Research Letters is different, with a more open policy on preprints and no embargo period on the accepted version (or the final version if open access). Applied Ecology and other Ecological Society of America journals allow preprint on the author or institution website, or a preprint server provided it is linked to the final publication, provided there is no peer-preview, and provided that the author can legally transfer copyright to ESA. I’m not sure exactly what ‘peer-review’ includes here (comments or more formal review)? And I assume that so long as the author retains copyright they can transfer it, though this perhaps ought to be further examined. But that at least is a taste of the complexities involved in this evolving field.

And don’t forget – these details may not be correct when you read this, so always refer to the respective authoritative sources before committing.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s