Over at Cosmic Variance, Julianne is annoyed at Nature's embargo policy. It seems that somebody or another posted a paper to the arxiv while submitting it to Nature, and included a note on the arxiv submission asking people to abide by Nature's embargo.
So, instead of blogging about the Incredibly Exciting Discovery (which I'd loooove to talk about), I'm writing about what a ridiculous fiction the authors are asking us all to participate in, for the sake of the authors' potentially getting a publication accepted to Nature. The authors advertised a paper to thousands of interesting, engaged scientists, who are then supposed to keep their mouths shut so that the authors can get a paper into a particular journal -- one that is not noticeably more influential in astrophysics (i.e. the difference between Nature and non-Nature is not nearly as big a deal as it is in biology).
The commenters are similarly down on Nature and the whole idea of submitting to Nature (conveniently ignoring the comment from an actual Nature editor saying that the authors have misunderstood the policy, and the embargo does not mean what they think it means).
What's really interesting here, though, is not what people are upset by, but what they aren't questioning, namely, the race to the arxiv. Everybody agrees that Nature's policy is bad because it seems to restrict publication and discussion on the arxiv, but nobody questions whether it's good to have scientists racing to be the first to post their preprints like idiot commenters playing "firsties."
While I think the embargo rules at Science and Nature are a little silly (though less silly than they used to be), I'm not convinced that they're the bad guys, here. Or, rather, I'm not convinced that they're bad in an absolute sense-- their rules may be worse than the arxiv-race culture, but I don't think either is good.
There's an obsession in science with the order of publication that I don't think is really healthy, and I think it's only gotten worse. At the Science21 meeting last fall, Paul Ginsparg talked about how there's a huge spike in arxiv submissions just after 4pm, because the daily update email puts papers in the order in which they were submitted, starting at 4pm. He said they can see scripts hitting the server to check the time, and then dumping papers in just as soon as the clock has ticked over. Apparently, the position of a paper in that email has a fairly significant effect on the number of views and citations that paper receives in the future.
That's just crazy, but it's a logical consequence of the race-to-publication culture that we have in academic science these days. Which is, itself, pretty crazy.
Science is not a short-term project. The real goal of science is the advancement of the sum total of human knowledge. The world is not likely to end in the few weeks between a paper being finished and the time it appears in Nature. There's no good reason for the sense of urgency you see in the comments on Julianne's article-- especially not for a paper on astronomy. It took millions of years for the light to reach your spectrograph-- you can wait a couple of weeks to talk about the results. I'm not even convinced that being first with some result actually does anybody all that much good-- I can think of several examples within my own field of AMO physics of people who weren't the first to publish on some topic, but are the ones who dominate the field, because of the quality of their science, not the timeliness of their publication.
But then, that's the culture of academia these days. People have a sense that if they're not first to the arxiv, or if they aren't discussing something the instant it first appears, they might as well not be doing science at all.
At some point, it would be great if everybody could take a step back, and a deep breath, and ask whether the academic science culture we have is actually healthy. That might cause somebody's preprint to be a few minutes late, though, so I won't be holding my breath.
- Log in to post comments
What's interesting to me about that post is the statement that "essentially all astronomers and physicists visit [ArXiv] daily." I'm a physicist and I never go to ArXiv, and I suspect no one else in my department does either. In fact, looking at the listings of papers in physics by topics, it seems that almost no one in any of the fields I work in (dynamical systems, nonlinear optics) uses ArXiv. I suspect that the obsession discussed here is primarily confined to smallish subsets of some particular subareas of physics.
Well said!
I agree with Chris that the arXiv phenomenon is limited to specific subfields of physics and astrophysics. That doesn't mean that other subfields don't sweat publication priority to an unhealthy extent, just that they don't go to the extremes that arXiv-intensive specialties do.
Obsessing about publication priority has always been an issue in science. As Tom Lehrer wrote in the early 1950s:
Hi Chad -- I'm not officially "annoyed at Nature's embargo policy". It is what it is, and if a group of authors want to go that route, that's fine and they understand the rules (which are actually more complex than I intimated). Instead, I'm bemused at being asked to participate it in it for the benefit of authors who simultaneously are advertising their work to a major scientific outlet (ArXiV).
Minor quibble aside, I absolutely agree with you about the "firsties" business. Next year, both the papers become Author1 et al (2009) and Author2 et al (2009), with no asterix pointing out that one of them showed up on ArXiV first. (However, I do have sympathy with junior scientists who feel compelled to do this because of the current horrors of the job market -- fear and scarcity make it hard to take the noble route sometimes).
Hi Chad -- I'm not officially "annoyed at Nature's embargo policy". It is what it is, and if a group of authors want to go that route, that's fine and they understand the rules (which are actually more complex than I intimated). Instead, I'm bemused at being asked to participate it in it for the benefit of authors who simultaneously are advertising their work to a major scientific outlet (ArXiV).
Minor quibble aside, I absolutely agree with you about the "firsties" business. Next year, both the papers become Author1 et al (2009) and Author2 et al (2009), with no asterix pointing out that one of them showed up on ArXiV first. (However, I do have sympathy with junior scientists who feel compelled to do this because of the current horrors of the job market -- fear and scarcity make it hard to take the noble route sometimes).
It's easy to dismiss this "firsties" business, but I don't think it's particularly new and I don't see any fundamentally different alternative. As long as the incentive system is based on achievement level, quantified in a necessarily imperfect way, we are stuck with all the flaws of such system. That's not to say that the system cannot be incrementally improved (say, by putting more emphasis on the refereeing process).
It's easy to dismiss this "firsties" business, but I don't think it's particularly new and I don't see any fundamentally different alternative. As long as the incentive system is based on achievement level, quantified in a necessarily imperfect way, we are stuck with all the flaws of such system. That's not to say that the system cannot be incrementally improved (say, by putting more emphasis on the refereeing process).
I think the "firsties" problem has gotten worse relatively recently. Certainly this embargo problem wouldn't've been an issue before the arxiv became dominant-- nobody would've been able to write about the paper early because nobody would've had the paper early.
I think the bigger problem is that the time scale for the scientific value of a paper to shake out is longer than the time scale on which faculty are evaluated. People feel like they have to play the "firsties" game because it may make the difference between high-and low-"impact factor" publication right now, because they're going to be evaluated on their research in a year or two, and need those high-impact publications right away. The fact that slower publication may well result in better science doesn't have a chance against the short-term incentives.
Doesn't the fact that, in this particular case, there were two papers being published by competing groups that were posted essentially simultaneously to the arxiv indicate that perhaps there was some coordination between the two groups, and that in fact there was some effort to avoid a mindless rush to publish first? It certainly appears that way to me...
I was in particle theory back in the early 90's, about when the arxiv got its start -- it was at that time a minor acceleration of the pre-existing preprint culture. Almost all of the papers that we read, we read as preprints, and even before the arxiv, most papers were pretty much out of date by the time they appeared in the journals. I don't think that the arxiv really increased the pressure for priority.
One way of rewarding the slower publication is using the hierarchy of journals. The publication with better scientific value might gather fewer citation and other immediate signs of recognition, but it will be the one published in the more prestigious journal, in recognition of its scientific value. Now, if only life could work that way...
I don't see evidence of an "arxiv race". They posted the same day, which means presumably they were aware of each other's work and they coordinated a joint release of their papers, as any decent people in such a situation do. What's to complain about? Are you suggesting they shouldn't have posted to arxiv at all until the paper is published? That's just a subfield-culture dependent convention.
Thought A:
If the Pons et al cold fusion paper had been sent to arXiv when first submitted, the job of the historians of science trying to sort out the Case of the Changing Figure and the Case of the Missing CoAuthor would have been a lot easier, and Jones et al would have known their agreement to co-publish in Nature had been broken.
Thought B:
The point of the embargo at Nature and Science (and Presidential speeches, etc) is to give journalists and bloggers a chance to chew over the paper before publishing, so their article will (hopefully) show more than a femtosecond of thought in the rush to "firsties" in the journalism world.
The point of tolerating arXiv in this environment is that it gives all scientists a fair shot at the next stage in the process, but with the fair warning that they might be basing their research effort on data that might not survive peer review. This is better in many respects than the situation Otto Frisch benefited from: you can scoop everyone on the measurement of the energy release from fission of U if you know about the discovery before the paper is even written.
Thought C:
Alex R @8 reminds me to ask if that rapid cycle where papers are out of date by the time they appear in print resulted in better or worse progress in theoretical particle physics? Were they out of date, or simply forgettable as Least Publishable Units?
I was in particle theory back in the early 90's, about when the arxiv got its start -- it was at that time a minor acceleration of the pre-existing preprint culture. Almost all of the papers that we read, we read as preprints, and even before the arxiv, most papers were pretty much out of date by the time they appeared in the journals. I don't think that the arxiv really increased the pressure for priority.
... in particle theory.
In fields that did not share the pre-existing pre-print culture of particle theory, but that have since taken to using the arxiv, it's a big change.
One way of rewarding the slower publication is using the hierarchy of journals. The publication with better scientific value might gather fewer citation and other immediate signs of recognition, but it will be the one published in the more prestigious journal, in recognition of its scientific value. Now, if only life could work that way...
It would be lovely, but as I'm sure you know, the real effect is usually in the other direction-- the second paper submitted to the more prestigious journal is more likely to run into trouble because it's not first, unless it's really remarkably better than the first paper submitted.
I don't see evidence of an "arxiv race". They posted the same day, which means presumably they were aware of each other's work and they coordinated a joint release of their papers, as any decent people in such a situation do. What's to complain about? Are you suggesting they shouldn't have posted to arxiv at all until the paper is published? That's just a subfield-culture dependent convention.
I'm talking not just about the specific papers Julianne was referring to-- I can't be, because I don't know what they were-- but also the culture suggested in the comments over there, where the prevailing opinion seemed to be that posting to the arxiv was vastly more important than posting in the journal, and if Nature wants things embargoed, people should not publish in Nature.
I think it's a little silly for Nature to insist on embargoes-- though, as was pointed out by an editor, the original note is apparently well beyond what Nature actually requires. At the same time, though, I think the idea that it's so important to get results out there quickly that you should forgo publishing in top journals in favor of quick release on the arxiv is insane.
I know from experience that Science turns things around very quickly-- the referee decision on our paper came in a matter of a couple of weeks, and that was over Christmas. It took a while longer to see print, but I think they held it up so they could publish it at the same time as a couple other papers on BEC stuff. When they've got a really hot result, they can get it into print really fast.
The notion that science can't wait a couple of months for peer review to run its course is just nuts.
Certainly this embargo problem wouldn't've been an issue before the arxiv became dominant-- nobody would've been able to write about the paper early because nobody would've had the paper early.
No. Before the arXiv, the authors would have sent out preprints to their hundred closest friends, each with the same note they attached to the arXiv posting. It wouldn't have been public, though.
I don't see the 4pm arXiv submission phenomenon as primarily a logical consequence of the race to publish, but as a reaction to shortcomings in information dissemination in the research environment. The way I understand it, the first arXiv postings of the day have a higher impact because of a lack of time or attention of the readers to absorb all of the day's postings. Similarly, the front page stories on a newspaper will be more widely read than those on page 11.
The general problem here is information management in an era of information proliferation. I think many researchers (in astrophysics at least) cannot or do not stay sufficiently up to date with other (published!) work of relevance to their own research. Since there is currently no one-stop shop to do so efficiently, authors will continue to pursue idiosyncratic means of improving the visibility of their research.
Everybody agrees that Nature's policy is bad because it seems to restrict publication and discussion on the arxiv, but nobody questions whether it's good to have scientists racing to be the first to post their preprints like idiot commenters playing "firsties."
On the other hand from the perspective of someone like me who has no chance of reading anything published in "Nature", this policy is pretty neat! I get to read everything as soon as it's written, whereas legitimate scientists reading legitimate peer-reviewed journals have to wait months. It's like I'm being rewarded for being an unproductive spectator.
HAHAHAHAH! Good one, holmes!