Announcing your findings (but not really).

Over at Cosmic Variance, Julianne Dalcanton describes a strategy for scientific communication that raises some interesting ethical issues:

Suppose you (and perhaps a competing team) had an incredibly exciting discovery that you wrote up and submitted to Nature.

Now suppose that you (and the competing team) simultaneously posted your (competing) papers to the ArXiv preprint server (which essentially all astronomers and physicists visit daily). But, suppose you then wrote in the comments "Submitted to Nature. Under press embargo".

In other words, you wrote the equivalent of "Well, we've submitted this to Nature, but they won't accept it or publish it if the news gets into the press, so can all of you reading this just not actually, you know, tell anyone? Oh, but can you make sure that you give us credit for the discovery, instead of the competing team? Thx!"

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 authors in this case are kind of announcing their findings to other scientists in their field -- but, owing to the embargo on their results, they kind of aren't.

What's going on here?

Let's start with the idea of a press embargo on results submitted to a journal.

One reason an embargo might be sensible is to allow time for a proper peer review of the submitted manuscript prior to publication. Scientists operate with the assumption that what gets published may differ significantly from what was originally submitted to the journal editors as a result of serious engagement between the authors and the peer reviewers. Done well, peer review unfolds like a serious conversation between scientists (albeit one conducted with written words rather than spoken ones). The authors of the paper are putting forward scientific claims and substantiation for them. The peer reviewers are giving the authors feedback as to whether they have made their case persuasively -- and if not, where the weaknesses in the argument are, and what might be done to make the argument more persuasive. The back and forth between the scientists on the submitting side and the scientists on the reviewing side is what makes the published results knowledge as far as the scientific community is concerned.

Of course, the press embargo here also serves Nature's interests. Nature would like to be the go-to journal for exciting scientific results, not just for scientists, but for journalists presenting science stories to a wider audience. Controlling the flow of information, and making sure what the journalists report on is the fully vetted and revised papers (rather than the unreviewed manuscripts) is one way Nature can further this interest.

But peer reviewed journals are not the only organ by which scientists can communicate their results to each other. The ArXiv preprint server is a mechanism by which astronomers and physicists of various sorts can communicate their results directly to their professional community. In doing so, they can engage in a more far-ranging, open-ended, and transparent (rather than confidential) process of peer review. They can get very rapid feedback about how persuasive a scientific case they're making, not from two to four of their scientific peers, but potentially from hundreds of them.

Uploading a preprint of your manuscript to ArXiv that includes a note indicating that the manuscript is under review at Nature (and thus subject to a press embargo) changes the equation. As Julianne describes it, the message is something like this:

We can't really talk about our results (because of the Nature embargo). But we want to let you know that we have these results. In other words, we want to stake our claim to priority for these findings, just in case someone else comes up with the same results while we let Nature's process run its course.

We will have something to talk about -- our findings are so cool and important that you're all going to want to talk about them -- but our interest in getting these results published in Nature trumps our desire to open this discussion right now.

Let's pull back for a minute to examine this interest the authors seem to be prioritizing. It could be that they have decided that, of all the scientific journals that publish in this discipline, Nature has the best editorial and peer review process; going through this process would then result in a higher quality paper. Or, it could be that the authors feel that Nature will reach more of the people they want to reach with these results than would an alternative journal without such restrictive embargo policies. Or, perhaps the authors anticipate that publishing in Nature will help them score more career-reward points than would publishing in an alternative journal.

Of course, ethics is a matter of how your interests and my interests come into contact. So it's fair to consider the interests of the other scientists in the community -- specifically of the members of the community who use the ArXiv preprint server.

The scientists in this community have an interest in the communication of good scientific information between scientists. They also have an interest in being able to work with and build off of bits of knowledge built by other scientists. They don't just want to know what you've discovered -- they want to use that to help them make discoveries of their own.

These other scientists don't, however, necessarily have a huge interest in helping you maximize your career-reward points. Most of them will refrain from blatantly sabotaging your quest for points by breaking the embargo, but given that the research they want and need to do to earn them career-reward points may hinge on being able to respond to your announced (but not really announced) findings, they will not look on you fondly.

Here, it's worth asking why, if you were the author in this case, wouldn't you submit your manuscript to Nature but refrain from uploading the preprint to ArXiv? It's much easier for the members of your scientific community not to break the press embargo on your cool new finding if they are blissfully unaware of it prior to its publication.

If the publication of the result in Nature earns significant career-reward points, why risk it by announcing the results on ArXiv? Either career-reward points aren't the only consideration on the table, or uploading the preprint to ArXiv itself garners some career-reward points.

Would ArXiv be useful in defending a priority claim? Could the judgment of the scientific community that uses ArXiv overrule Nature's judgment on the matter?

There seems also to be some ambiguity about what would count as breaking the press embargo. Presumably, if conversations between scientists about the findings were happening in real life, with no reporters in the room, it would not violate the embargo. Having such discussions online might be counted differently because of the persistent pixel-trail of the discussion -- something members of the press might find. But, as Julianne suggests, members of the press might as easily find the preprint in ArXiv themselves and break the news (and the embargo) to further their own interests.

Finally, there is the very curious feature of this particular case: two groups of authors uploading preprints to ArXiv on what looks to be the same finding, both with the note that their manuscript is under review by Nature and thus that it is under a press embargo. One wonders if the two groups were aware of each other (working in the same area, closing in on the same finding) prior to the deposit of the first of the manuscripts with ArXiv, or prior to the submission of the first of the manuscripts to Nature.

What level of competition is there really between the two teams? Do they have the same results (and the same interpretation of them), locking them into a fight to get through peer review and get published first? Do they have the same results but interestingly different interpretations of them, making their results complementary and both potentially publishable? Do they have contradictory interpretations of the same data, meaning that at least one group will be judged wrong? (What about the possibility of sharing authorship (and priority) in the manner of Sean Cutler's experiment?)

One would think, in the second and third of these circumstances, that a vigorous discussion among the ArXiV readers would be useful -- at least, it would be useful in furthering the community's goal of building a reliable body of scientific knowledge. To the extent that a journal's embargo policy preempts this kind of vigorous discussion among experts in a field, it can be seen as undermining this goal.

But the score-keeping does play a role within the scientific community, especially in terms of how scarce resources (like research funding, permanent positions, and tenure) are distributed among scientists. Practically, the members of the two competing teams here might argue that they need an important publication in a high-impact journal even to stay in the scientific game. If they prioritize the community's interests, they might not get to remain as members of that community.

And here, perhaps the community of astronomers and physicists who are regular ArXiv visitors might want to start thinking about whether collectively they have the clout to change the game at least a little.

If this situation ultimately comes down to the question of who gets across the finish line first as recognized by Nature's publication date, it ends up being both boring and frustrating in terms of how the interests of individual scientists seem to undercut the interests of the scientific community to which they all belong. Why, given the choice, would a scientific community engaged in vigorous online discussions about scientific findings give Nature the authority to make the crucial judgments?