Simply Statistics A statistics blog by Rafa Irizarry, Roger Peng, and Jeff Leek

Authorship conventions

The main role of academics is the creation of knowledge. In science, publications are the main venue by which we share our accomplishments, our ideas. Not surprisingly, publications are heavily weighted in hires and promotions. But with multiple author papers how do we know how much each author contributed?  Here are some related links from Science and Nature and below I share some thoughts specific to Applied Statistics.

It is common for theoretical statisticians to publish solo papers. For these it is clear who takes the credit for the idea. In contrast, applied statisticians typically include various authors. Examples include the postdoc that did most the work, the graduate student that helped, the programmer that wrote associated software, and the biologists that created the data. So what position do we assign ourself so that those that evaluate us know our role? Many of us working with lab scientists have adopted their convention: the main knowledge creator, usually the lab head, goes last and is the corresponding author. Here are examples from Jeff, Hongkai, Ingo, and myself. Through conversations with senior Biostatistics and Statistics faculty I have been surprised to learn that many are not aware of this. In some cases they went as far as advising junior faculty to publish more first author papers. This is somewhat concerning because junior faculty could be faced with study sections (where our grants are evaluated) that look for last author papers. Study section is not going to change so I am hoping this post will help educate the statistical community about the meaning of last author papers for those of us working in genomics and other lab-science related fields. Here is a summary of authorship conventions in these fields:

  • Last and corresponding author is associated with the major contributor of ideas and leadership. This is the most desirable position.
  • First author is associated with the person who did most the implementation and computing work. Very good for a postdoc or jr faculty. Excellent for a graduate student.
  • First and corresponding is sometimes used when the person not only had the ideas, but also did half or more of the work. This is rare.
  • Big collaborative projects will have two or more corresponding authors and two or more “first” authors. I included an example above.