On the future of scientific communication

Dipterist extraordinaire David Yeates writes:

If accepted, a recently proposed amendment to the ICZN allows for electronic publication of taxonomic names…. [T]he logical implications of this proposal are many and far reaching. For example, this change may lead to further advances so that zoological taxonomy bypasses traditional journal publication entirely

I agree with Yeates.  Taxonomy will migrate from paper journals to online databases, and this will happen sooner rather than later.  But I think it worth noting that this is reflective of a broader change in scientific communication already underway.

Scientific papers are fossils, relicts from the era when writing on paper was the only medium to reliably disseminate data and ideas.   Now that we have other forms of transferring information, there isn’t any reason to continue packaging all scientific output into periodical two-dimensional prints. Especially not when much of it (e.g., simulations, data visualization) is particularly ill-suited to paper.

Here’s a not-very-risky prediction.  Within the next decade, papers will be winnowed to something resembling press releases or commentaries.  The experiments themselves- the data, methodological details, and statistical algorithms- will inhabit a network of databases.  Users will be able to access original data,  replicate statistical analyses, or add their own data or try novel sets of analyses.  It will be a more flexible and transparent system.

We are already taking baby steps in that direction.  Raw data and methods are increasingly presented as online supplements.  Sequence data is already stored in GenBank.  Online portals (like CIPRES) already allow for remote analyses.   PLoS has been experimenting with open commenting and review systems.  Once universal data standards are adopted and the infrastructure kinks that prevent greater integration are worked out, I imagine the migration will happen quickly.

A few hurdles to the post-paper world remain, however.  Chief among these is the academic rewards system.  Hiring and promotion decisions for scientists currently depend on the frequency and stature of traditional publications in a pathologically myopic way (leading to, among other perversities, the division of single studies into unnecessary series of smallest publishable units).  Working scientists will have little incentive to contribute to emerging infrastructures until career metrics are reformed to better reflect other kinds of productivity such as data generation and creation of software and other tools.

10 thoughts on “On the future of scientific communication”

  1. Thanks for the heads up. The new regulations seem good, except that I think ZooBank registration should be compulsary also for paper-based publications – after all, these are in my experience the ones most readily forgotten in diversity assessments, electronic databases etc.

  2. Oh great. Not that I don’t favor more transparent, and rapid publishing and increasing availability of raw data, however, I was hoping the dawning age of Data Rapers (e.g. macro”ecologists”) would be a passing phase. This virtually guarantees that hordes of “scientists” will now build entire careers out of writing review papers and doing “meta-analyses” of old and existing data. Why bother doing your own research and experiments when someone else can do it for you? For that matter why bother knowing what organisms look like or do when you only need to go online and harvest their sequence data, apply a name, and read a press release somebody else wrote about what their biology is all about. It couldn’t be easier to be a scientist! Heck I might even try it.

    1. Your comment touches on the point of my post. We need to reform the career performance metrics we use to gauge scientific output. If researchers earned credit for generating data and contributing it to databases like genbank, EoL, and antweb, then the incentives favoring “Data Rapers” (I’ve never heard that term before!) would diminish.

      But I’m not sure I agree. It’s true there are more meta-studies around than before. But it’s also true that there is a lot more data available as well, and I’d be surprised if the rise of meta-studies has surpressed the generation of new data rather than simply added a level of analyses that wasn’t possible before (the merits of which are certainly debatable, of course).

      I’m more concerned about the loss of academic positions for field and organismal biologists overall resulting from the increasing reliance of Universities on the overhead generated by Big Science grants.

  3. Alex,
    I totally agree with you, both on the need for more “ology” positions, and the need for revised career metrics. However, I would counter that with the formation of entire institutes (e.g. NCEAS) dedicated to “meta anaylyses,” where you cannot even participate as a post-doc or researcher if your proposal includes field work (!) suggests that my fear is already being realized. Furthermore, take a look at the placement of post-docs into the “plum” ecol/evolution positions from that institution over the past few years versus the more traditional post-doctoral model and I think you might be as discouraged as I am. By the way, as far as I know the term “Data Raper” is my own invention and it really does apply.

  4. I don’t see papers as fossils. The constraint to a certain number of pages and a certain format actually forces many people to actually THINK how to transmit the acquired knowledge rather than just dumping raw data somewhere and selling this with a few nifty press-release style paragraphs.

    Have you ever looked at “supplementary material”? Even on journals like science the writing quality (and thus comprehensivesness) often drops massively in these appendices. And how exactly do you except anyone to do a peer-review on what you propose? I certainly would hate to dig through heaps of data with little explanation. If you do put the full information, you are basically back to a “paper”.

    It is sure that printed journals are practically replaced by online abstract + PDF. Moreover, it is absolutely no problem to add more colorful pictures or simulation movies online. However, the difference is that any paper should, no, MUST be understandable also without this information.

    Finally: how exactly do you think “we” can or will change the way scientific output is measured? If anything, the competition will get more fierce. I am, unfortunately, very pessimistic about that.

    Hopefully some processes change and will be improved to speed up the whole thing. But writing a paper to me is much more than being constrained to a medium. It is a process by which you think and distill what is actually relevant, important, and meaningful.

    1. I look at “supplemental material” all the time. Unfortunately, the most important parts of papers (the original data) are often shunted there. As publishers don’t think much about it, they don’t enforce writing standards nor do they screen the material for problems. As you point out, that’s a serious issue.

      But I’m not arguing for an amplification of the current system of journal + supplemental. I’m arguing for a more far-reaching change where online data, methods, and tests assume primary importance. The emerging primacy of databases should see a corresponding increase in quality-control and peer-review. Just because the current guardians of online publication do a poor job of enforcing standards does not mean that all online material should be forever sloppy.

      Re: peer-review. Too much of what passes for peer-review is limited to proofreading and relatively minor cosmetic fixes in the writing. At least, this is true of my field (systematics). As long as reviewers lack access to raw data and implementations of statistical tests, the review process will remain poor at uncovering serious flaws in the way data are handled and interpreted. As scientific communication and the associated data and analytical algorithms move online, reviewers will be able to not just examine the methods but *replicate* them. That’s a powerful improvement, in my opinion.

      Re: changing the measures of scientific output. I wish I knew an answer to that- I’m not sure that we can, but I think that we should.

  5. Pingback: Tweetlinks, 10-13-09 [A Blog Around The Clock] « Technology Blogs

  6. James C. Trager

    I don’t feel particularly suited to comment on this, but I wrote my son (a rising organismal biologist who also has recent experience with meta-analysis – in press). He says:

    “As to Josh’s comments about meta-analysis, I can attest to the fact that a well-conceived and thorough m-a is not easier than most field-based investigations. Different, certainly, but not easier. Also, I am certain that as the novelty of the approach wears off, reviewers will be more critical of meta-analyses. Case in point: not so long ago, anything using GIS or new genetic methods could get published, even if it was shit. However, as those approaches became more commonplace and the average reviewer’s knowledge of them became more sophisticated, simply including certain methodologies was no longer enough to merit publication. As quantitative reviews become more common, I suspect the same will occur.”

Leave a Reply