Sep 1, 2018
Preprinting in a Material World
Scholarly communication is changing. Well, it’s always changing. As for example, it is only recently that peer review was only viewed as the “gold standard” and gatekeeper of the scientific process.
Preprints are part of the next step in the evolution (or re-evolution?) of scholarly communication. To help others, I’m going to present a why/what/how/when and risks piece on the matter.
Why Preprints are important?
Unfortunately, researchers are increasingly busy — the volume of scientific communication is increasing exponentially (see Figure 1). This increase in quantity, regardless of quality, can cause significant headaches for the peer review system. We are asked to review more work, this often reduces quality, and the system is struggling.
Regularly editors are having to ask for 10+ people to review before they get a single acceptance, and then they have to nudge and persuade the reviewers to complete their volunteer task in a timely fashion.
Publication of academic work in journal papers is slow and cumbersome, and the reward timeline can be difficult to manage, especially for vulnerable groups (e.g. early career academics and job hunters).
Within my group, it can take someone 3 months to write the manuscript. 1 month for internal sign off. 1–2 months of industrial liaison (if needed). Then we enter the lottery of journal publication processes and peer review, where the flood of submissions and the variability in peer review can be problematic.
This means that it typically takes us between 2 months (fastest I can think of) and 2 years to share our work in the public domain, enables others to critique and build upon it, and also provide credit of the work undertaken within the group.
What are preprints
Preprints are archival snapshots of work, often supplied to a significant and recognised system. Materials science does not have a discipline specific preprint server, so I tend to publish work to Condensed Matter — Materials Science on the Arxiv.
These preprints are open access, free to download, searchable, and indexed by google scholar. The ArXiv is funded by a consortium of partners (mainly Cornell, the Simons Foundation, and 206 member institutions). The ArXiv has been operating since 1991 originally at Los Alamos National Lab and it moved the www.arxiv.org domain in 2001 when principal operations moved to Cornell.
There are many other preprint servers in existence — including the wonderful EarthArXiv which was founded by a group of geoscientists including Prof Chris Jackson and others.
How — the process of preprinting
You can preprint a manuscript whenever you want. Different fields have different cultural attitudes towards the medium, but many items in the ArXiv actively have revisions. These revisions are not erratum, they are simply improvements to the work. Some of these improvements will come from informal peer reviews (just email the author of the work with a comment) and others may be the result of formal peer review at a journal.
Once I have the pdf (if writing in Word) or tex documents gathered, it takes me about five minutes to complete the preprint submission. Usually, the submission is made available online the next day and after this announcement I am free to share the preprint link with colleagues, post on my website, and tweet about the work.
Once the preprint has been announced I am sent an email with a password that I send to my co-authors and they can claim ownership to add the paper to their profile.
When to preprint
We preprint when we have something significant to say. Once the article is announced this does not mean it has been reviewed or that it enters the traditional scientific canon. However, the work can be cited, read in multiple timezones, read without active distribution to a 3rd party (e.g. as a submission bundle for a job or grant), and it can be built upon and worked into the next great idea.
Frankly, the act of preparing a manuscript for full peer review (via the conventional route) and concurrently submitting it to the ArXiv means it has likely had more attention than most modern conference proceedings papers.
However, you may wish to preprint at a later time. When I first started preprinting, I only released work to the ArXiv that were “accepted articles” (be careful about embargoes) which were of the same format as submitted to my institutional repository (Spiral).
Recently, we have found that one of the greatest benefits in the ArXiv model is that we can cite work according to our own timeline. After the article is submitted and announced (often 24 hrs), I can cite my own work in the next piece of work or a conference talk. My co-authors can cite this work, as preprint material, in their next job applications.
Risks
There are several risks I think about when I consider preprinting:
1. Have we made a mistake?
Yes, we probably have somewhere in our work. We’ve also made mistakes in our peer reviewed work. We made it better and we published the next thing.
Each time we make mistakes, we learn a lot in the process, and we were never intentionally dishonest in publishing something we have thought was wrong at the time of publication.
It can be embarrassing, but publishing a revision is a whole lot less embarrassing than the feel of a correction paper, cunningly disguised in some Latin as an “erratum” or “corrigendum” (though I have only just found out that an erratum is an error introduced by the publisher & a corrigendum is an correction to the work by the authors).
2. Will we get scooped?
Preprints let you control when items are published. You can fix your narrative and timeline. You can avoid the games-person-ship that people play when they hold up articles in review so that they can polish off their own work. Not all of your work will be fit or suitable to preprint (e.g. commercial limits and industrial liason can be difficult) but at least you can drive your own story more often.
Formally, your preprint on the ArXiv post has an announcement date. This date is fixed and is linked to the date when the original article is submitted. It will not get changed or rewritten when the article is revised, or when the peer reviewed copy hops from one journal to another. This makes a preprint more robust to avoid scooping than other forms of peer review publications.
In terms of prior art timelines, the ArXiv timeline is view-able by the whole human community of researchers. The date and content of the release are archived.
Now there is a risk that we will end up getting “scooped” in terms of presenting an idea that is used to feed into another idea that makes a greater story. However, this risk is apparent when you are ‘just’ networking over a beer at a conference with your frenemies.
3. Copyright / embargoes / legalease / media
Make sure you can release the work and you have agreement of all your authors to preprint.
You do risk breaking a story early and reduce the ability to control/hype it’s impact, but I believe that if you are publicly funded then fame and fortune should play second fiddle to the core idea of releasing your know-how and findings back to the people who fund you. Furthermore, there are some very interesting media embargo policies (like this from the Royal Society) as it is unlikely that you will be first person in your predicament.
Many of the journal publishing houses accept that preprints are in the fabric of scholarly communication and support archiving and sharing of work as preprints (see Nature, Elsevier, Springer).
The statistics from Sherpa/Romeo show that there is a favourable amount of support for preprinting (Figure 2).
If you are worried about preprinting, just ask your journal editor. Or you can search the SHERPA/RoMEO database and pick a journal that lets you self archive. You can also lobby your previous favourite to change its mind.
Final word on PrePrints
I don’t think there is a true final word. The publication space is changing. Preprints are likely just a step in our journey forward.
Within our group, it is not likley that we will preprint everything we are involved in for a few years. There are a complex range of factors that go into how we manage our publication strategies. We will try to preprint more, and our approach to preprints will likely evolve, as I hope will the Materials Science and Engineering community.
For the next steps, I suspect we’ll also start getting involved in curating an overlay collection or two. One significant concern with preprints is that we could end up flooding the market, if we cut out all the gate keepers for publication, but I argue that this is happening already with pay-per-publish light-touch-peer-review journals from major publishing houses. At least the ArXiv makes this process less expensive, more community focused and supports the evolution of a story.
At Materials Scientists and Engineers, I hope we can step forward together to embrace the open science movement and support preprints more widely. They offers support for our early career colleagues, enables our industrial partners to improve their uptake of key ideas, enables dissemination of our research to emerging research markets, and will likely enable us to move the discipline forward together more effectively.
This article was motivated in part by one of my favourite Canadians, Dr Sinclair. Thank you.
If you like this post, please ‘clap’ it below to recommend it to others.
If your are interested in my research group head to: http://expmicromech.com
If you fancy getting in touch, telling me your stories, and helping me understand the world better, you can find me on twitter as @BMatB, or keep up to date with the group’s work via @ExpMicroMech.
