Monday, January 05, 2009


Author lists on peer-reviewed papers

PhysioProf notes (rather rudely, but I gather that’s his style) his opinion on the order of authors listed on multi-author papers:

The second and third authors are asterisked as “equal contributors”! What a motherfucking joke. The only people in the entire world who could possibly give a shit about that are their parents. HAHAHAHAHAH!
He does note that his comments apply only to papers in his field, and that other fields might have different customs.

Indeed; this is a tricky point. It’s easy in many cases (and some of the comments to PhysioProf’s post point out that when it’s not easy, it’s often due to bad project management). Whoever did the most work on the project gets to be first author. If there’s a faculty advisor or other senior person who has ultimate responsibility, she goes last. Everyone in between, they say, is chopped liver.

For the rest of this, I’ll talk about my field, computer science.

So let’s back up, first, and talk about who gets to be an author. In the Real World, someone who writes a significant part of the text is an author. Anyone who doesn’t, isn’t. But this isn’t the Real World; it’s Academia. Here, an “author” is someone who contributed significant work to the project, someone who was responsible for some of the results that the paper is describing. It helps if that person also wrote some of the paper, but it’s not strictly necessary (and, depending upon that researcher’s writing skills, the other authors might be very happy about that). It’s considered Very Bad Form for some project members to get together and write a paper, leaving one or two project members out.

But there are two ways to list the authors, and it’s not always clear which method’s been used.

Method 1: The first author is, ideally, the person who contributed the most to the project and to the paper. Other authors are listed in order to significance of contribution, with equal contributions or minor authors placed in alphabetical order. There can be problems here: suppose Smith did the most work on the project, but Jones took the lead in writing the paper, and led the way to doing extra work to get more detailed results needed for a good journal article. Is it Smith or Jones who gets listed first?

Sometimes decisions like that will come down to who’s more politically important, who’s in greater need of more first-author papers, or something like that. Sometimes, they’ll be considered “equal contributors”, and we have the “two first authors” situation for which PhysioProf expresses severe disdain.

Method 2: List all authors strictly alphabetically. This can have odd consequences, giving a so-so contributor first billing because his name happens to be Aanders (or relegating Zelazny to the end of the list, even though she wrote 70% of the text and had the original idea that sparked the work).

But the fact is that being first is important, and especially so if there are more than two authors. A paper by Aanders and Zelazny will probably be cited as “Aanders and Zelazny”. One by Aanders, Jones, and Zelazny will likely be cited as “Aanders, et al”, and if there are more authors the “et al” citation is a certainty. No notations in the paper about levels of contribution will change that.

But what about decisions for hiring, tenure, and the like? Yes, in computer science, as in biomedical science, you need a selection of first-author papers to make the right impression. If Aanders and Zelazny collaborate frequently, they’d better take it in turns to be listed first... no one is going to give Zelazny a break because her name is at the wrong end of the alphabet,

This makes author lists an unfortunately cutthroat game, sometimes. I’m happy that it’s never been a point of contention in any papers I’ve been involved with.

No comments: