For us scientists, publishing our work in a respected scientific journal is one of our top priorities. To do this, it must pass the gates of peer review. . . The idea behind peer review is simple: Step 1) editor filters out garbage papers that aren’t worth the reviewers’ time; Step 2) send the paper to expert scientists in the field to identify strengths, flaws and overall worthiness; Step 3) reject, accept, or return the paper to the authors for corrections. This process of getting expert feedback on each other’s work is a good idea at its core, but some things about this system are broken. Mainly, it often takes FOREVER. Also, most journals aren’t interested in publishing negative results or repeated work, and at the end of it all, the articles may still not be publicly available.
Open access journals – peer reviewed journals that do not charge subscription fees – have recently gained popularity. It’s a nice option, but many of the most highly regarded journals charge the authors a hefty fee in order to make their paper open access… presumably to off-set what the journals might have been making in subscriptions (although I highly doubt that the numbers add up). To give you an idea, our lab recently paid over $6,000 just to make one of my papers open access. Thankfully we could afford it, but these exorbitant fees mean that although a researcher may want to publish open access, many times we still choose the cheaper option.
Another movement is now gaining traction, though: publishing on preprint servers. Many big-name journals (Nature, Science and PNAS, to name a few) are now on board with the idea of submitting papers to preprint servers before submitting to the journals themselves. This has the huge benefit of making the papers available to everyone – researchers and public included – virtually immediately (think: 1 day). Granted, this system is not officially peer reviewed, but other researchers can still comment on flaws and the authors can submit revisions – In a way, the system lets the paper reach far more ‘reviewers’ than conventional peer review would. An added benefit is that if it really is quality stuff, then other researchers can integrate that knowledge into their own work without waiting the 3-6+ months for a peer reviewed journal to publish it.
It’s no wonder what motivated this movement: publication bias is a huge problem in many disciplines (science is no exception), and preprint publication offers one way to counteract that. The root of publication bias is that journals (and scientists, for that matter) want to publish new and provoking work – of course, that’s what’s most interesting to read. But this means that it is extremely hard to publish work that replicates previous findings (how boring; been there, done that!) as well as negative results (nothing to see here!). Maybe counter intuitively, these kinds of results are actually very important to disseminate.
Consider the following situation: 10 different researchers do the same experiment independently, working on a new form of chemotherapy in mice. One researcher finds that the treated lab mice had significantly smaller tumors than untreated mice – everyone else found that there was no difference. Which experiment do you think will get published? The first one, of course! Even though the evidence overwhelmingly points to this just being due to random chance. If this becomes a published paper, it is unlikely that it will be uncovered as such because a) no one wants to publish an experiment that’s already been done and b) negative results are not very interesting. And so it lives on, accepted as truth. This is what makes preprint servers so useful: with the publication selection bias removed, the barrier to publish replicated work and negative results is very low. This allows bigger picture patterns of the data from independent groups to be revealed, when before it was shrouded with bias.
I recently submitted a paper of my own to a preprint server (BioRxiv) and have had a very positive experience. First, the process is fast and easy – as an added bonus, BioRxiv allows direct submissions from their server to many prominent journals, without wasting time with reformatting. Second, it is visually pleasing – the title and abstract of the article are easily searchable and is formatted nicely on the BioRxiv site. BioRxiv also takes advantage of social media by showing a real-time twitter feed immediately below the abstract, listing all tweets directly related to the article (what a confidence boost!). Finally, each article is given a DOI, making them easily citable in your future papers.
I should note that this whole idea is nothing new: a similar preprint server (arXiv) for math, physics and most other quantitative disciplines has been up and running for over 20 years and has been highly successful. For some reason, biology and chemistry have lagged far behind, with BioRxiv launched in 2013 and ChemRXiv coming soon. It’s time to step up our game. As long as your journal of choice accepts submissions that have already been published preprint, I highly recommend this service to help combat the peer review woes.