As some of you might realize, I have been in a pretty shitty mood for most of the last week (though there were a few spurts of happiness as well, so at least I've had my moments). It's been more of a general feeling of being down rather than, say, rantiness (that ought to be a word, even though it appears not to be one) or something more extreme like bitterness. However, a small bout of mild depression (without any specific cause, by the way, so it's likely that I just need to do better with sleeping, eating, and getting people to distract me to get back to my version of normal) doesn't mean that I am not going to try to satisfy my deep-seeded desire (need?) to rant:
Today my collaborator Peter Mucha e-mailed me to pass along Nature's new authorship policies.
Previously, Nature "strongly encouraged" statements of what each author did. Now they are following the practice of other journals---an increasing number of them, I think?---and requiring the listing of such contribution statements. I object to this rather strongly for a very simple reason: One of its big effects is to increase the potential for bad blood between people. There can be a ton of intra-group strife in science as it stands, and as far as I can tell, enforcing things like this exacerbates the problem. When I submit to a journal that requires such statements, I grudgingly include them, but whenever it is optional, I have a very strong preference for not including them. People's understanding of their own papers (and knowledge of the contents therein) will be judged when they give a talk on the bloody thing (or otherwise demonstrate it firsthand)!
To be fair to Nature---though since when was I ever fair to nature?---I should mention that there are other things in the new policies that I support. Those do take the form of box-checking, but the ideas behind what checking the box is supposed to represent are ones I support. I just have major problems with the enforcement of policies with so much potential to add to jealousy and intra-group strife. What is supposed to be gained by doing this? I know my opinion on this matter is far from universal and is reasonably likely to be in the minority (maybe even a rather small one?), but frankly I just don't get it. Nature wants to make author contributions more explicit, but to what end? If somebody is out to screw a coauthor, one can just game that anyway, so I don't see how it can discourage those kinds of hideous practices. Oh, one other thing I should mention on Nature's behalf: They do seem to be more flexible than some of the other journals in the format of contribution listing. (Some of the biology journals, for example, seem to be so experimentally oriented that they do not include any fields that allow one to give author contributions for things like "designed algorithms" or "wrote code". One is only allowed to use the categories they provide, and at least in some cases, this misses some very fundamental ways in which authors might have contributed!)
It's not realistic for things to be changed in the various scientific fields, but when it comes to paper coauthorship, there is one thing that I think the pure mathematician got exactly right: Always make it alphabetical. (And then one's knowledge can be judged when giving presentations, etc.)
Grumpily Yours,
Mason
I would include "designed algorithms" and "wrote code" as either "designed, implemented, troubleshot experiments", or "analyzed data and interpreted results", depending on the purpose of the code. Of course there is always "conceived of study", "drafted manuscript" or "critically revised manuscript".
ReplyDeleteWe did something like the first two you mentioned as the 'best fit' (actually, I think for the journal I am thinking of, those two are encompassed in the same category anyway), but it really doesn't need to fit into any of those categories. I don't see how code or algorithms has anything to do with conceiving of the study or anything whatsoever at the manuscript stage. The whole thing is a bunch of bs.
ReplyDeleteof course it is BS. they just want there to be a paper trail of accountability.
ReplyDeleteThat's ridiculously anal. I'm glad I just deal with normal astronomy journals rather than Nature. My current paper is at the nearly-final-draft stage where we start thinking about the author list. There's clearly a case for cutting a bunch of people since their only contribution was being on the original proposal five years ago, or for some of the younger team members a few minimal comments on an earlier draft. But as you say there's lots of potential for bad blood if one takes an anal, Nature-style approach. I've had someone bitch to me about being lumped into "and the {group acronym} team" rather than being mentioned by name in another team member's talk, even.
ReplyDeleteReally, Nth author papers don't count for a whole heck of a lot, it's just the first author and to a limited extent the next few (before the shift to alphabetical order) who actually matter. May as well keep feathers unruffled and be generous with coauthorship.
Being generous with things like coauthorships and acknowledgements is always the best policy.
ReplyDeleteJustin: One other thing that will matter in many fields is the last author because that person will be assumed to be the group leader.
The only journals to which I have submitted that have enforced this are bioinformatics journals and the really general ones like Nature. However, while the origin appears (to my eyes) to be from the biological sciences, the practice seems to be spreading.
Huh, that last author convention most certainly does not apply in astronomy. I don't really assume anything about the group leader - sometimes I know because a certain person is famous for doing the kind of thing in the paper (or I know the group and know who the leader is directly), otherwise it could be anyone. And it doesn't really matter, that I can think of. If I want to ask for something related to the paper (input files so I can recreate one of their figures and add my own data points, say) I'll either ask someone on the author list who I know, or I'll ask the first author. Whether they're the group leader or not doesn't make much difference.
ReplyDeleteIn our group, there's no distinct leader, either. Depending on how you count, there are 5-7 people who PI'ed successful observing proposals (depends whether you count some fairly small proposals). I'd characterize four of them as "group leaders" to some extent. Their PI-ness puts them quite early in the author list on relevant papers, not last...
Good to know about all this. Hopefully if the US astro journals start making noises in this direction, we can make a stink about it. Probably not much the US community could do if Monthly Notices of the Royal Astronomical Society gets it in their head to try this, though. :-)
The physics labs are almost always run by a single PI, and theory groups in physics are often run that way as well. I believe that this is also true for biology, chemistry, etc.
ReplyDelete