I had just submitted a manuscript to Environmental Science and Technology, when not 30 minutes later I received an email saying it had been “unsubmitted” for being too long. Now, this manuscript, at ~5100 words, wasn’t so wordy. It was the figures, which just happened to be large due to the nature of the work (resource sheds, which I will be posting about soon). Apparently the figures are assigned word count equivalent values, which are used to estimate the length of the paper. This left me with a choice: shorten the front material or go elsewhere.
The solution was to move the one table I had to the Supplemental Material, and make some hard choices about how to divide the panels of my figures between figures to be presented in the printed article, and figures for the supplement. That is, I didn’t want to part with anything. Which brings me to online material. There was a time when online material meant data dumps. But now it seems more and more journals are relying on supplemental online material for material that used to be acceptable in the main article. I find this to be a double-edged paring knife.
On the one hand shorter length limits cull verbose writing, which is always a good thing. On the other hand, not all complete studies are necessarily short nor should be divided. One could argue that there are journals for longer works. But it’s the higher “impact” journals that have tighter length limits. So, large studies that cannot fit into small papers must be excluded from high-impact journals? Ah, the devil’s advocate responds, you should be publishing in a variety of journals that are individually suited to the material. Well, yes, but tell that to a struggling graduate student or postdoc looking for a job. The pressure is on to publish in high-impact journals because such publications do, in fact, make a difference.
Further, this is changing the way we write and read papers. Science Magazine, for example, generally pushes methods to the online material. Pamela Hines, Senior Editor at Science, pointed out that because of this practice, Science papers are of comparable total length to publications elsewhere when all supplemental material is considered. That is, the fun stuff (motivations, context, results, conclusions, implications) is presented up front. Oh, and if you really want to know how the study was done, then you can read the nuts and bolts in the supplemental materials. Dr. Hines asked me if I wanted to see more Brevia in Science, which takes this partitioning to the extreme. I said yes, but only because it would mean more could get published, and hence would create more chances to get in. But if more papers were presented as Brevia, and not just those studies that fit the venue well the nature of reading papers would radically change.
Yet I see us moving in that direction anyway. After all, the cost of publishing is dropping faster than the economy with increased reliance on electronic distribution (for what percentage of your papers do you actually buy paper reprints for anymore?). And given the ever cheapening cost of computer memory, there really are no practical limits on manuscript size. Which is, of course, why the use of supplemental material is increasing.
Funny, I had never really thought about supplemental material for my own work until I had to for my last two manuscripts, both submitted in the same week to different journals. My Science Brevia submittal required the use of supplemental material from the start, and consists of the methods. The ES&T manuscript, in contrast, presents ancillary results. But those results only became ancillary when I was forced to make cuts. I’ll be interested to see how the reviewers and the editor regard the partitioning. Ironically, I think the ES&T manuscript is better for having been shortened. This also means that for every paper I write henceforth I will consider supplemental material as a standard section, or rather, have to.
If the trend of supplemental material continues, we could see a time when all papers get longer, but only when all elements are re-assembled. Or better-yet, lets make “papers” dynamic files embedded with all the information the author wants to present, set to display the way the reader wants to read.