A question recently piqued my interest — in general, how much of a journal’s “submission to publication” process is consumed by ultimately favorable editorial and peer review? That is, if it takes 100 days to get an acceptable paper through to publication, what proportion of that time is consumed by the editorial review and peer review process that ends up accepting that paper?
This doesn’t address the review that leads to rejection and resubmission, but only editorial and peer review that leads to publication. How much of the time from submission to publication does favorable peer review require on average?
Examining submission, acceptance, and publication dates from a dozen journals — 11 from Nature and, as an external benchmark, PLOS One — comprising more than 6,000 papers across a variety of fields, business models (subscription and Gold OA), and architectures (traditional journals and megajournals), it looks like favorable editorial and peer review consumes on average 80-85% of the time from submission to publication.
There is a wide range of normal, of course, from five days to 1,698 days, and from 10% to 98% of the total time to publication. But averaging everything together, the range of 80-85% is where these review processes settle.
This is something I didn’t know before, and it seems to have a bearing on efforts to speed up publication cycles. Editorial and peer review is something authors and readers value and expect. Reviewers are volunteers, are harder and harder to find, and aren’t under the direct control of publishers and editors like production or administrative staff, making it less likely efficiencies will be derived from this part of the publication process.
This also indicates, in a more precise manner than experience and intuition, why posting papers on preprint servers is so comparatively quick, while publishing editorially reviewed and externally peer-reviewed papers takes time — the vast majority of the time a paper spends with a journal is spent with editors and peer-reviewers.
The 11 Nature journals I had queried for submission, acceptance, and publication dates were:
- Nature (flagship)
- Nature Genetics
- Nature Medicine
- Nature Immunology
- Nature Biotech
- Nature Materials
- Nature Neuroscience
- Nature Physics
- Nature Geoscience
- Scientific Reports
- Nature Communications
I also had PLOS One queried because it is a high-volume journal with an approach to review that suggests the process might run more quickly. Simply measuring across the Nature family of journals might have reflected the way Nature operates, how their editorial and peer review systems and software work, how their policies work, and so forth. PLOS One seemed far enough away in many respects to serve as a solid check, and a potential clincher if a common finding emerged.
The biggest surprise in the data was that even for PLOS One, which has a reputation for peer review based mainly on a paper being deemed “scientifically valid,” the time between submission and acceptance consumed 85% of the span between submission and publication.
As expected, PLOS One was the fastest on average to publish, clocking in at an average 189 days from submission to publication, with Scientific Reports coming in second, at an average of 222 days. PLOS One was also faster on the production side, taking an average of 24 days to publish a paper after acceptance, with Nature Communications coming in second at 33 days and Scientific Reports in third, at 35 days from acceptance to publication.
Every journal could be quick when warranted, as well. Nature’s fastest time to acceptance was less than 2 weeks, and some papers went through production in a week. PLOS One had some papers get through peer review in as little as 5 days.
These weren’t unexpected findings, as speed is a variable for publishing, and one that’s not entirely controllable for a variety of reasons. Also, a paper that takes longer to publish may or may not be superb, just as for any paper published quickly. There’s no correlation I’ve ever seen between publication speed and quality.
But what caught my eye was the consistency in the percentage of time peer review and editorial review consumes, no matter the amount of time on the proverbial clock.
- Nature — 80%
- Nature Genetics — 82%
- Nature Medicine — 77%
- Nature Immunology — 83%
- Nature Biotech — 85%
- Nature Materials — 85%
- Nature Neuroscience — 78%
- Nature Physics — 81%
- Nature Geoscience — 81%
- Scientific Reports — 82%
- Nature Communications — 83%
- PLOS One — 85%
There are a few additional angles to this.
These times don’t encompass the time (and money) spent on rejected papers. Rejecting papers is definitely an expensive process. When I analyzed the cost of rejecting papers for a journal I used to run, rejection accounted for 69% of the expense of handling papers, with the editors and publishers having nothing really to show for it. The remaining 31% of the cost was spent getting manuscripts to the point of acceptance. The costs in both cases are generally fixed expenses — manuscript software, salaries and benefits, stipends to editors, and facilities costs.
Of course, production systems keep getting more efficient. Or do they? It’s tempting to think that perhaps peer review is consuming more time to publication because it’s now so much easier to push articles online. Yet, I haven’t met anyone doing cartwheels about how easy online publication systems have become to use, run, and manage. So, barring compelling evidence otherwise, let’s call this one a wash.
Peer review takes time, costs money, and delivers what the market expects. Most journals can make it happen quickly at times, and sometimes it takes years to get a paper through a journal’s peer review process. Judging from these data, it appears a good rule of thumb is that peer reviews takes 80-85% of the time a manuscript spends in-house prior to publication.
Does that match your experience and process?