Monday, October 9, 2017

The secret lives of manuscripts

Everyone who has written and published a scientific paper has pondered the mystery: you send in your final version, then there’s this long silence. Suddenly you get an urgent email with your page proofs, insisting that you go through them with a fine-toothed comb and get them back within two business days.  “Hurry up and wait“, I’ve heard multiple authors remark, rolling their eyes.  For most scientists (myself included), our publishers are a cryptic “black box”. We know something is happening to our paper, behind the scenes. But we’re not really sure what happens out there,  why it takes so long, or why there’s suddenly such a rush when the proofs appear. Okay, the black box metaphor is boring and overused. How about we think of it as the mysterious years after young salmon (our paper) venture out into the wide ocean, out of sight, before suddenly returning to their natal stream for spawning (proofreading). Um, that’s not exactly a great metaphor either, but I’ll swim with it. So here’s a blog on the unseen part of your manuscripts’ life history.
            Last week I got to venture under the publishing ocean surface, visiting the journal office of The American Naturalist at the University of Chicago Press (UCP). I’m gradually transitioning into the role of Editor-In-Chief of AmNat. I officially start January 1 2018, but am ramping up my activity by starting to share some Editor tasks with Judie Bronstein, the outgoing editor. I visited the Press for a day to learn more about how the journal works, meet the people who make it happen, and talk about ideas for the future.

I started my day in Chicago with a visit to a Hyde Park institution, Valois, where Obama was a regular. I got to sit at the Presidential table. Then, well fed, it was off to the University of Chicago Press.

            When I walked into the UCP board room, the first thing that struck me was how many people were there.  I had expected to be meeting with Trish Morse (the public face of the journal at scientific meetings), Owen Cook (who works closely with authors during revisions and preparing final version), and Valerie Bajorat (the Publisher, who I’d corresponded with), and maybe a couple of other people. But the room had a dozen people in it, and they quickly apologized on behalf of a few people who were out sick or traveling. I was genuinely surprised that so many people had come. Over the next few hours, they took turns explaining to me what each person did, answering my questions, and charting the unseen stages of a manuscript’s  life-cycle.  By the time I surfaced from this dive into the publishing underworld, I had a much greater appreciation for the value-added that a good journal office provides. And by extension, a greater appreciation for why publishing costs what it does, and why that is worthwhile.
To pass on what I’ve learned, let’s track a hypothetical AmNat manuscript from submission onwards.  Let’s start with the part that is at least mostly familiar to authors, though perhaps not in detail. The first parts below will be familiar to most authors, though maybe not in the level of detail, the number of steps. That’s what I want to convey though, that there are many, many steps:
1. The first thing a manuscript encounters as it leaves its natal stream (your computer) is the Editorial Manager website. This is a shockingly complicated (but flexible) commercial system that the journal subscribes to (which costs money). The AmNat Editors and Staff have personalized many aspects of the system over years, building in an informative but imposing set of visual flags, messaging systems, auto-alerts, reporting tools. There is of course a staff member (Rob Blixt) dedicated to keeping this system operational, and optimizing it to make submission as quick as possible for you, but as informative as possible for us.
Owen Cook (left) and Rob Blixt (right)

2. Once your manuscript is submitted, the Managing Editor (Trish Morse) or Owen Cook check to make sure the basic requirements have been met. Unlike some journals, we don’t require a specific format for review (again, to make submission as easy as possible for you; though be forewarned that reviewers often get agitated if they think your paper doesn’t match journal style). If the paper passes this check, it gets moved into a folder where the triumvirate of Editors can see it.

Image result for Trish Morse
Trish Morse, who many readers may have met at conferences.

3. One of the three editors will claim a paper (sometimes after a bit of haggling amongst ourselves), then read it over. We may opt to send it back to the authors with a clear justification for why it isn’t suitable. These can be several-page reviews by the Editor; we don’t want to make such a decision lightly or arbitrarily. If we think the paper has a chance, we will check the list of Associate Editors: who is suitable to handle this paper’s subject, setting aside people who are unavailable, or already handling a full load of papers. We then post on Editorial Manager a list of the AEs we think would be good, and hope they take it. If not, we revisit the list of names.
4. The Associate Editor then looks over the paper to decide whether it is worth reviewing. We don’t want to waste authors’ time if their paper doesn’t have a chance (poor fit for the journal, or too clearly flawed), nor waste reviewers’ time. If the paper seems worthwhile, then the AE proposes a list of names of reviewers (see for some comments on this process). This involves looking through the manuscript’s references, searching on scholarly databases or Google for people with appropriate expertise, checking for conflicts of interest, and perhaps tracking down contact information if the reviewer isn’t in the Editorial Manager database. The proposed list (often  ~6 names) goes back to Trish Morse at the journal office. She or Owen checks the list for conflict of interest and availability, since AEs don’t always have the time. invite reviewers and process the responses until we hit the targeted 2 reviewers. Often, they go back to the AE to get more names until we hit the two-reviewer target. Unlike some journals, which blast out all the email invitations at once (thereby often getting an excess of reviews, which uses everyone’s time), AmNat sends out invitations until we get exactly two reviewers. That takes just a little more time, but it is better citizenship, I believe, to not draw on too many peoples’ time.
5. AmNat gives reviewers 21 days to review.  They are busy people, after all, and volunteering. Some journals demand faster reviews, but we want to give them enough time to do a careful job. Rushed reviews can be sloppy (missing a mistake, or misunderstanding a point) and cursory and are more likely to be grumpy. That Editorial Manager website proves its worth again with review reminder reports with the flexibility to respond individually to reviewer.
6. Once both reviews are in, they are checked for completion and problems by Trish or Owen and then routed to the Associate Editor, who typically reads the paper a second time and writes a substantive review in their own right. One of the things that we pride ourselves about at AmNat is that the AEs really work to fix any flaws that the reviewers happened to miss. Many of our AEs also go to great lengths to identify diamonds-in-the-rough; manuscripts that are flawed but contain the ingredients for a great paper. Some will take a paper through multiple rounds of revision, providing detailed feedback on writing, graphics, and pitch until the paper meets our standards. My favorite example of this is detailed in Meghan Duffy’s blog post about how our (then) AE Yannis Michalakis helped her ms (previously rejected at Ecology) improve until it won the ESA’s Mercer Award ( .
7. The AE sends the reviews and their own recommendation to the Editor, who typically reads the paper again, and adds their own insights as well in a decision letter. There are often Skype calls or email conversations at this stage to negotiate a mutually agreeable decision. Last week alone I ended up in three separate conversations with AEs who wanted some feedback on their recommendation. Once they submit a recommendation, my decision letter has to be written, then sent back to the Managing Editor (Trish) who checks it, formats it, and sends it. I usually give the paper another read-through before writing my decision. Speaking as an author, I can’t over-emphasize how valuable the feedback from AmNat’s Editors can be. Yannis Michalakis has (as Editor) greatly improved several of my papers with feedback that went above and beyond what the reviewers and Associate Editors provided (themselves giving very good feedback).
8. For each round of revision, the paper may or may not go out to reviewers (usually only  if the changes are substantial and the AE not able to evaluate themselves), but will be read by the AE and often the Editor, until it is clear that the paper is good enough to publish, or clear that the paper isn’t on a trajectory to reach that level.  Then a final decision is rendered, which brings up format issues for a smooth journey through Production,and the author makes any last changes and submits a “final” version to the office via Editorial Manager.

Let’s pause here. Everything up to this point is moderately familiar to scientists. But let’s put it in perspective. At the time the paper is submitted in its “final” form, it has been handled by Trish Morse or Owen Cook somewhere between 6 and a dozen times. There have been dozens of separate steps on Editorial Manager. The Editor has read the manuscript two or more times. The Associate Editor has read it two or more times. The two reviewers have each read it at least once, often twice. So it has been read through, on average, about nine times (assuming two rounds of review), and has involved six people. But so far, a large fraction of those people are volunteers (AEs, reviewers).  This is the point at which the manuscript leaves the natal stream and enters that unseen world.

9. Picking up where we left off: the submitted “final” version is checked over in detail by Owen Cook, to make sure all the parts are present and accounted for. Data is deposited, files are complete, appendices and supplements sorted out. Owen makes sure that figures conform to journal standards for size, font, resolution, and are converted into vectored EPS format for the highest quality online and in print. He’ll communicate with authors to fix any remaining problems, and clarify what is destined for print appendices versus online supplements.  Forms and agreements are gathered. When files, figures, and forms are complete in the journal office, the paper gets assigned to an issue (usually the next one headed to Production). Then the paper is sent from Editorial Manager to the UCP Production system. This process takes as long as it takes authors to bring the paper up to Production standards.

10. Once in Production, the Production Coordinator Jeannie Harrell checks it into the Production database, makes sure all the necessary metadata and elements have come through from the journal office. the paper has to be translated to a new file format (XML) for the next stages. To do this,  Jeannie sends it to a professional service. This way we can get LaTex and Microsoft Word and Open Office files into a single publication-ready version.  The service returns the paper within 48 hours, where it is checked for the necessary tagging and code by the Publishing Specialist and Production Editor Samanatha Tansino.
Jeannie Harrell

11. When she gets the file ok from Sam Tansino, Jeannie delivers the manuscript to the copy-editing team (“Editorial” led by Mary Nell Hoover).  The manuscript PDF now is printed on paper as the last word on the authors’ intentions, placed in a green folder, and brought downstairs. Editorial goes through the paper line by line, looking at grammar and phrasing. They convert your equations into XML-friendly format for the clearest presentation. They check that things are capitalized, or not. They check citations in the text and reference details at the end of the paper. They check that you are consistent in using symbols or abbreviations in the same way. In short, they go over the paper with the proverbial fine-toothed comb, with the Chicago Manual of Style close at hand on every copy-editor’s desk. Their goals are to ensure the writing is clear, readable, and conforms to journal style. As needed, they will change sentences, capitalization, punctuation (adding proof queries to the author about any changes in meaning). The editorial team also formats tables into journal style. They even check a lot of our numbers in tables, for instance making sure that sums add up. Just to get an AmNat manuscript ready for the quality control check can take an editor up to 36 working hours, depending on the size and complexity of a manuscript. Then the other team member -checks it, sending queries to the author as needed. Because our in-house copy editors are working on many papers (for many journals) at once, this can take roughly 20 days.

Mary Nell Hoover

12. When the copy editor is done, they hand the paper over to another team member for checking, When those changes are made and approved, the paper is sent to the typesetter, who has two days to return the proofs. These are not immediately sent to the author. Instead, the Editorial team checks the “pre-proofs,” especially the math, and either approves them or sends them back to the typesetter for corrections.  At this point the copy editor also makes sure all queries to the author are ready. This editing-typesetting-checking process is repeated until the copy editing team thinks the proofs are ready. All told, this step usually takes about 5 days more.

13. This is when the proofs get sent back to you, with a request that you return the proofs within two days with any last (minor) corrections.  Two days always seems rather demanding to authors, who are unaware of what’s been going on in the background, but as you have seen the paper has passed through many hands, many times.  And it’s not done. Problems with returning the proofs can be solved, but the deadline is to keep the issue as a whole on its monthly schedule.

14. When you return the pdf with corrections, it is received by the Production Controller, who sends it downstairs again to the copy editing team. They check your responses, which is at least the fourth time they are seeing your paper. Then the paper is sent back to the typesetter for re-typesetting; they have one day to return it.

15. When re-typesetting is done, office staff get a notification and download the zipped version. They print out a copy of the typeset pdf on paper, and send it back down to the copy editing team for one last check. All told, every single character in the math in every article has been checked against the author’s PDF at least three times. When they are satisfied the paper has no errors, it is returned and marked as finalized and ready to be posted online. At this point it goes back to the Publishing Specialist, who puts the HTML  version and the typeset pdf on the journal website. Now, your paper is posted as an Ahead Of Print (AOP) article. The digital version and its green folder are put in a pile to await a complete issue’s-worth of articles.

16. The articles then have to be sorted into an order by the Editor In Chief.  The articles must then be paginated, leaving room for advertisements, announcements, editorials, and the like, to form a complete issue. The whole issue is printed on the same kind of paper that forms the actual journal issue, for one last review by the copy editing team. Mary Nell makes corrections to the whole issue, sends it back to the typesetter if any corrections are needed. This cycle is repeated as many times as necessary to get a version that is approved by the UCP Editors.

17. A print-ready pdf is made by the type-setter for the whole issue, and sent to the printer. The printer returns a digital and a hard-copy version to be checked one last time, printed now on the same paper stock as a real issue.

18. After a last check-through, the issue is approved, sent to the printer, and published as an electronic edition and e-book, as well as printed and mailed to libraries and subscribers by the UCP Distribution Center.

19 Only after the paper is safely tucked into its issue does the Billing Manager take the page charge calculations from the Managing Editor to work out the invoices because the charges are tied to actual pages (unless an author requires a flat fee article processing charge)

Now, this team of people cannot operate without some other support. There are the electronic publishing experts, who solve the tough tagging and presentation issues. There’s the IT people who make sure the Editorial Manager database and website, and journal website, are operating smoothly. And fixing people’s desktop computers as needed. There’s the marketing staff who keep subscriptions coming in from institutions, and seek new institutions to work with. There’s human resources staff who pay everyone. There’s the janitorial staff, maintenance. Many of these people of course help with the whole University of Chicago Press, which handles a moderately large number of journals.  

So, why does this take so long?  Because careful publication takes time. The University of Chicago Press believes that scientific publications should be as accurate, readable, and professional as possible. As you’ve seen, that takes an incredible amount of behind-the scenes work to make sure that everything from the tables to figures to grammar to copyright permissions are perfect. Your paper passes through many people’s hands, with many iterations of corrections. The end result is a higher-quality product with fewer mistakes.  As an author, I’ve long been impressed with the detail and professionalism of the AmNat copy editing team, which finds many small details to query and correct, often far more than other journals pick up. Better still, let’s contrast this with PLoS One (not to name names), which makes authors do all their own copy editing, and doesn’t even do a round of proofs. Let’s face it, most of us just aren’t trained as copy editors. As a result, I’ve found PLoS One papers to be full of stylistic flaws, typos, and errors that a professional copy editor would catch.

Why does it cost a few thousand dollars per article (few authors shoulder that whole cost)? The income generated from our roughly 120 articles per year (and past papers) brings in the income that keeps these people employed to help your articles be as clean as they can be. The income also has to pay to use the Editorial Manager software, and contribute a bit to keeping the lights and water on in the UCP as a whole. Those page charges you pay are a very modest contribution towards supporting that behind the scenes staff, but by no means covers all the costs. The balance comes from institutional subscriptions, and individual subscriptions, and society memberships. Despite UCP’s reliance on subscription income, it gives away free journal access to institutions in over 100 developing countries, about 8,000 universities in total.

The end result is that AmNat produces a very high quality product, even though it is actually one of the cheapest journals to publish in or subscribe to.  A moderate number of authors pay nothing at all. That’s very important especially for early career researchers who may not have the financial resources to cover even regular page charges. Those who do pay regular page charges are covering a moderate fraction of the costs of producing their article. People who opt for open access pay various higher rates depending on the level of access. At the extreme, we offer the opportunity to cover the full costs of production for extreme levels of open copyright.

So when you balk at a bill for page charges, remember the hard work of the large team of people behind the scenes who are laboring to make your paper into a high-quality product. In the case of The American Naturalist, the journal is a not-for-profit (501c3) entity. The University of Chicago Press is a branch of the University of Chicago. Its primary task is not maximizing income for investors or an owner, but promoting academic pursuits. The building and offices are clean and well maintained, but not fancy.  I asked for directions to the room where you can roll around in big piles of cash, and the staff looked confused.

Occasionally on Twitter I read comments by people who basically want to shift to an all-BioRXiv publishing model. We self-publish, and “get rid of journals”. Cheaper. Faster. No annoying peer-review setting standards (just post-publication review).  Personally, I’m not on board with this.  I think the review process gives great value added, and that’s backed up by a totally unscientific poll I did on twitter: a vast majority agreed that reviews improve papers slightly (50%), or very substantially (40% of 88 votes).

So if ever someone advocates getting rid of journals, I have my canned response. GOOD journals can:
1)   help improve your paper through anonymous (and thus more frank) reviews

2)   improve your paper through Associate Editor and Editor comments that seek to bring out the best in your paper, or direct you to a journal where your paper will most readily reach its target audience

3)   help you produce a polished and professional final product that reads well, is easy to understand, and looks good. As a result it will be read more, and cited more.

4)   Distribute your paper to readers via their website, table of contents, subscriptions, and social media.

5)   Apply any excess income back to academic societies, which in turn support student research, travel, conferences, and the like, building a richer academic community.
These represent value-added to the scientific enterprise. That value added isn’t free, because it takes digital and personnel resources. So who pays? It either has to be the government (good luck asking Trump for that), the author, or the reader(s). Each has its flaws. Government payment is subject to political interference. Author-payment creates a barrier to entry for underfunded (especially junior) scientists, which hampers their career. Reader payment reduces readership access and citation (though remember we give away AmNat for free to thousands of institutions).

To conclude, repeat after me:

Journals provide value-added.

That value-added has a cost associated with it, which someone must pay.

Support your society journal (especially if not published by a big lucrative conglomerate).

Non-profit and open-access are not synonymous things.

The American Naturalist is awesome (though I am admittedly biased), both because of the great authors who submit interesting papers, the Editors and Associate Editors and Reviewers who work to improve those papers, and, let’s never forget, a large and hard-working editorial team that makes the high-quality final product.

A huge thanks to the staff of The American Naturalist at the University of Chicago Press, for the in-depth tour and education.

1 comment:

  1. Neat post, and I’ll raise a toast to University of Chicago Press’s and American Naturalist’s fine “artisinal, small-batch approach to scientific publishing” as I saw it called over at DE. It’s not an approach that can work for all society published journals, but certainly produces a fine product and makes the case well to “support your society journal (especially if not published by a big lucrative conglomerate).”
    However, one “drawback” of the American Naturalist hands-on, in house copyediting approach may be that they might less likely to publish new species through spellchecker taxonomy and evolution. This is where the big publishing conglomerates with their offshore or multi-discipline copyeditors can shine. For instance, “Chironomus dilutus” is an aquatic insect. Through the modern miracle of spellchecking and autocorrection, “C. dilutus” was transmorgrified in at least 80 published articles to “Chironomus dilutes.” The latter is a two-word sentence, rather than a species binomial.
    Curious, American Naturalist was not among the list of journals that I found accelerating the pace of evolution through automated copy editing.

    More at: