|
|
|
|
|
|
|
|
|||||
| Home Help Feedback Subscriptions Archive Search Table of Contents | |||||
First published online June 17, 2009
doi: 10.1242/10.1242/jcs.054825
| ||||||||||||||||||||||||||||||||||||||||||||||||||
Sticky Wicket |
|
Pop quiz. How much does a scientific paper cost? If you started tallying page and color-figure charges, you aren't even close (five points from Gryffindor). Here's a clue: how much does a lab cost to run per year? Make sure to include all salaries, lab space, lights and telephone, and depreciation on equipment. (Okay, that's being a bit of an accountant – you can probably get away with adding up all the grants, additional salaries and support, and grant overhead, and we'll call it a ballpark figure. For those of you without ballparks, we're talking about rough estimates, which I could have just said, but just be happy I'm not saying `seat-of-the-pants estimates', whatever they are.) We're referring here to labs that have only one type of product (papers); if your lab also produces things industrial, maybe ask someone in the other type of lab, assuming they will tell you. They might, and you'll say, `Gee, I didn't know you could run a lab on so little...'.
Now, take the total cost for a year and divide it by the total number of papers that were published from your lab that year. I know, the papers were worked on during other years, but it evens out. This will yield cost-per-paper. But Mole, you say, we don't only produce papers – we generate knowledge and train researchers and think about, um, stuff. Well, the knowledge and thinking have no value at all (as product, of course, it has value otherwise) unless it is communicated. And all that training doesn't generate an actual scientist, unless they are experiencing the ins-outs-trauma-pain-and-joy of publishing. So there: cost-per-paper. If you've never looked at it this way, it's a shocker.
Some labs are very economical, and produce papers that cost only tens of thousands of dollars (please convert to your favorite currency), but these are the exception. One-hundred thousand dollars per paper is still thrifty. Generally, I find that most papers that are published in reasonable journals cost about two- to five-times as much, by this calculation. (I'm privy to a lot of cost-per-lab and cost-per-paper details when I review grants in different countries – and the high end of my estimates holds pretty well throughout the first world, anyway.) Of course, I'm thinking about biomedical papers, and we're not counting the mega-big science papers, just the average sort of paper that comes out of a lab. All papers are expensive, except, possibly, one I remember in one of those weekly journals with the nice soft pages, concerning a `fairy ring' of mushrooms that had grown up where a dead animal had been noted the year before. That paper had no supplemental figures (and this may be its first citation, except that it isn't really a citation), which, in case you are just now joining us, is what we are talking about – supplemental figures, tables, and thingies that get tacked electronically onto our papers and kick around in a sort of cyberspace wasteland hoping to be glanced at.
There seems to be a lot of grumbling going on about the ever-increasing demand to publish supplemental figures and tables that are only available electronically, together with the hard-won data that makes its way into the `real' publication. But because the supplemental figures and tables require a bit of scrounging around to find, they are rarely seen by anyone. Many of us see this as a growing problem, adding to our difficulties in both producing and reading acceptable publications. But I, being the Mole, see this as a slightly different problem. No, I do not think that it is a bad thing to go from DNS (which, as you remember, is data not shown, or quite likely data not showable) to `here it is if you'd care to have a look'. But the fact that even interested readers often don't really care to have a look is part of the real problem. Which is this: it is so easy to demand that incremental bits of often irrelevant information must be obtained (`must' being implied by reviewers and editors), and this raises the very high-stratospheric costs of a paper into low orbit. I'm sure I could launch a satellite into low orbit for the cost of some papers.
When a reviewer asks that an experiment that has been performed on one foot in response to that reviewers' previous request now be performed on the other foot, it is just another supplemental figure, so neither reviewers nor editors generally hesitate to request it. So do it we must, and nobody actually cares, except, of course, the bench monkey (sorry, I meant researcher) who has to do it. And me. And you. Because this is all costing an awful lot of money. Fifty-six supplemental figures might finally get the frickin' thing published (and good riddance), but maybe we could have put that money towards something a little more important? Our supplemental data doesn't cost anything to put online (okay, it does cost something, but it lives on the electronic equivalent of cabbage water), but it costs a very great deal to generate.
It's time to liberate Supplemental Siberia. Close it down, flush it out, or, at least, use it for what it should be for, and not what it has become.
Here's my simple idea. A start on a way to perhaps turn things around, something to discuss. Lets say we put a moratorium on supplemental figures. We allow supplemental methods, which do take up a lot of limited space and are valuable for folks who want to repeat the work, check the details and, more likely, actually use some of the approaches in their own work. And we allow things that can't be published on paper, like movies, sounds, large resolution copies of photos in the main text and long tables of array data and the like. But we are going to say a big `no' to additional-experiment figures in the supplemental materials. If it is really important, it should go in the main body of the paper.
This is not really meant as a restriction on authors (although it may help to focus the papers that are coming out). This is a message to reviewers and editors that every wish for more, more, more should require the sacrifice of something that is already in there. If something, like a control, is woefully lacking, then yes it should be done in an experiment that replaces the one that was deficient.
For this to work, it is likely that we're going to need some negotiation. A step that will come between review and revision. It will have to include editors and, often, the reviewers. Really, this isn't such a big deal. Authors will suggest which experiments will be performed as really improving the paper and, when agreed upon, the necessary work will be done. Yes, we'll all act like this is a huge tax on our already precious time, but really, it is nothing compared to the time it will take our hapless authors to experimentally address every idle whim and wish. And it will save a lot of money.
Perhaps most of all, it will give the decision-makers a chance to take another look before sending the troops back into the breach. A chance to gain some fresh perspective and re-evaluate the reviews, which were probably done in a hurry anyway. True, sometimes the perspective will be negative (or more negative) – this paper really isn't appropriate for this journal and, rather than asking the researchers to collectively bang their heads against immovable objections, they would do well to move on. But often, I think, after a breather, reviewers might be urged to pick the one or two things that the authors should fix to make the paper actually, significantly better, and not just bigger.
But what will we do with all the data that now won't appear anywhere, even in Supplemental Siberia? Well, you can still publish it, say, in review articles to illustrate a point and shore up the conclusions. Or, if you really love it, set it up on your own website (wuh.wuh.wuh.MolesData.org). But let's bring the best, most important things back from the frigid electronic wilderness and give them a place in the sun. Lets direct our collective efforts towards the pursuit of bigger questions at a lower cost-per-paper, and answer things we feel we really need to know.
Here's a wild suggestion, whether you agree or not – talk about this. Maybe even send a note (or this column – hey, they might like the cartoon) to colleagues and editors you know. Refine this sketchy idea and see what might work.
Because the point is, this is our literature. The public pay for it, and they deserve that it should be as readable, accessible, economical and timely as we can manage. We can do better.
Dear Mole,
Since you're asking, I should tell you that JCS discourages the publication of supplementary material. And it's worth noting that it's not free – currently, our monthly costs to store supplementary material are $550. (For an idea of how this amount will grow over time, have a look at supplementary material figure S5 from last issue's Sticky Wicket!) Although we realise that some data – in particular, movies and complex datasets – cannot appear in the main text, we cap the amount of such supplemental material at 5 MB per paper. However, this cap is unpopular with some of our authors (particularly if they wish to include data such as multiple movies of 75 MB each)!
Best wishes,
Fiona Watt, Editor-in-Chief
Dear Fiona,
Aha, so even Siberia has its hidden costs!
Love,
Mole
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||
