Authorship Tiers

I recently led the writing of a report on software sustainability and high-energy physics: https://arxiv.org/abs/2010.05102. This report effectively has three groups of authors, with each group listed in alphabetic order: Katz – Stewart, the organizers of the workshop, leaders of the analysis and summarization, and lead authors of the report; Assamagan – Sexton-Kennedy, speakers in the workshop and authors of sections of the report; and Evans – Tunnell, contributors to the workshop and to the report. Note that there were also other workshop attendees who contributed to discussions in the meeting but are not authors.

I would have liked a way to have indicated this in the heading/authorship section of the paper. I know I could have indicated this in the text somewhere, or perhaps via many footnotes, but these didn’t seem like good options. I also could have used CRediT (Contributor Roles Taxonomy) but it seemed like overkill for this purpose. Additionally, CRediT defines roles but doesn’t explain how the authors value those roles, which I did want to do here.

In one of my other activities, I’m an Associate Editor-in-Chief for the Journal of Open Source Software (JOSS), which in effect means that I act as the Editor-in-Chief for issues related to specific papers one week out of every five (handling new submissions, reviewing them for scope, assigning them to editors, working on any problem, and performing the final acceptance of publishing of submissions that complete review.) One interesting paper I accepted and published this week is:

Hoffer et al., (2020). Minerva: a light-weight, narrative image browser for multiplexed tissue images. Journal of Open Source Software, 5(54), 2579, https://doi.org/10.21105/joss.02579

Note that this paper uses footnotes to indicate that the first two authors should be considered co-first authors, and that the last author is the corresponding author, which typically in this position indicates the PI or leader of the group/team. Again CRediT (Contributor Roles Taxonomy) could have been used here but might have been overkill.

An interesting related idea in economics is the use of ⓡ between author names, which indicates that these names are in random order, and that the authors made equal contributions. (The American Economic Association has published both an article and a policy statement about this idea.)

I wonder if this could be expanded on to define authorship tiers? Rather than ⓡ, we would use ① between randomized author names in the first tier, ② between randomized author names in the second tier, etc. Combined with CrediT, this would allow both the explanation of author roles as well as a indication of their value to the author.

Published by:

danielskatz

Assistant Director for Scientific Software and Applications at NCSA, Research Associate Professor in CS, ECE, and the iSchool at the University of Illinois Urbana-Champaign; works on systems and tools (aka cyberinfrastructure) and policy related to computational and data-enabled research, primarily in science and engineering

Categories UncategorizedTags 4 Comments

4 thoughts on “Authorship Tiers”

  1. Hi Dan, a common, simple situation might be a graduate student doing much of the hands-on work for a paper, guided by their PhD advisor. How do you such a situation should be tiered?

    Another thought, how about percentage credit? You can divide it up however you like as long as it adds to 100%?

    CRediT seems to be about defining roles in the creative process rather than weighing what each role contributes. Obviously, different people (author groups) might weigh different roles differently. Seems like there may be room for both.

    But I can also see all of this as time-consuming and potentially divisive. So how far do we *need* to go down this quantification rabbit hole?

    Like

    1. Hi David,

      Re the student and advisor, this seems like something that CredIT handles well.

      Re percentage credit, I explored that idea a bit in 2014 and 2015 in http://doi.org/10.5334/jors.be and http://doi.org/10.5334/jors.by

      And yes, I agree there’s room for multiple things, and this is a rabbit hole.

      I wrote this because of my two experiences this week, where in both cases, it was clear to me that there’s a gap and I had an idea to help fill it (the gap, not the rabbit hole)

      Like

  2. I like the idea of inline random tiering in the way you propose. We had faced a similar issue in https://doi.org/10.12688/f1000research.23224.1, where we had a group of primary authors who have conducted the original research and have written the bulk of the text, and a second group of authors who have contributed in a less comprehensive way. In the end, we have internally ordered both groups alphabetically by last name, and have put the first group first in the list while also marking all individual primary authors with a footnote noting equal contribution. A way to mark randomization such as the one you propose would’ve helped avoid the semblance of implicit order in either of the groups, as alphabetical order across the whole list seemingly doesn’t exist.

    Like

    1. Notably, in our case there weren’t any internal hierarchical differences in either of the groups, so we didn’t face any of the issues that David Bernholdt describes in their comment.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s