Random coauthor order could increase fairness and encourage collaboration

In a new NBER working paper, Debraj Ray (Columbia) and Arthur Robson (Simon Fraser) propose an alternative to the standard practice in economics of listing coauthors in alphabetical order. Let’s start randomly assigning coauthor order. Queen’s economist Christopher Cotton discusses this proposal.

By Christopher Cotton, Queen’s University

In economics, we almost always list coauthors alphabetically on a research paper. Only occasionally do we see coauthors list their names in some other order, a choice which typically implies very unequal contributions by the authors.

This field-wide practice has a number of downsides. For example, it unfairly gives greater visibility to coauthors with last names earlier in the alphabet. The standard citation practice of writing only the first author’s name, followed by the latin phrase “Et al.” (translated “and others”) means that first author enjoys greater visibly, even if in economics being first doesn’t imply one did more work. Just ask my coauthors in Cotton Et al. (2016).

Additionally, the current practice makes it difficult to give higher credit to one of the authors, if the author who deserves higher credit has a name which comes first anyway. For example, one can be pretty sure that Hicks should get more than half of the credit for the work in Hicks and Cotton (2016) (the paper was based on his excellent MA essay). But, giving Hicks extra credit was only possible because Hicks comes after Cotton in the alphabet.

In contrast, the author order in Boleslavsky and Cotton (2015) would be the same if we wanted to signal equal contribution, as it would be if we wanted to give more credit to Boleslavsky. It is most likely that we both provided an equal contribution. But, the reader can’t be sure. The only thing the reader can be sure of is the authors don’t want to give the majority of credit to Cotton.

The standard practice can discourage collaboration. A project team may hesitate to add another coauthor, if that new coauthor’s name comes first in the alphabet, even if adding the new collaborator could improve the quality of a project. People may think twice before adding a coauthor who is will reduce their own visibility in citations.

What’s the alternative to the alphabetical standard? One option is to switch to the standard in the sciences, where authors are listed according to the importance of their contribution. (Actually, the practice is more complicated than this, as it is often more prestigious to be listed last than in the middle of a long list of coauthors.) I don’t particularly like this practice, mostly because I’ve heard stories from colleagues in the sciences about the tensions that arise and friendships that have been lost while trying to determine author order.

Additionally, I imagine that the majority of economics papers involve coauthors who make similar-sized contributions (partially because, unlike in the sciences, we tend to only list the most significant collaborators as coauthors, relegating RAs, and research team members to an acknowledgment). Because of this, a default assumption of equal contribution makes sense.

In a new NBER working paper, Debraj Ray (Columbia) and Arthur Robson (Simon Fraser) propose a really promising alternative: Let’s start randomly assigning coauthor order. The abstract of “Certified Random: A new order for co-authorship” is below:

In economics, alphabetical name order is the baseline norm for joint publications. A growing literature suggests, however, that alphabetical order confers uneven benefits on the first author. This paper introduces and studies certified random order, which involves randomization of names that is institutionally ratified by a commonly understood symbol. Certified random order maintains all the ethical niceties of alphabetical order, but in addition: (a) it distributes the psychological and perceptual weight given to first authorship evenly over the alphabet, (b) it allows either author to signal credit when contributions are extremely unequal, (c) it will be willingly adopted in a decentralized manner even in an environment where alphabetical order is dominant, (d) it is robust to deviations, (e) it dominates alphabetical order on the grounds of ex-ante efficiency, and (f) barring the addition of a simple symbol, it is no more complex than the old system, and brings perfect symmetry to joint authorship.

Certification of random ordering is an important component of this. No group of coauthors will unanimously agree to go random if readers are unlikely to recognize that the order was in fact random. The fear is that a reader will infer that a non-alphabetical ordering signals that the fourth author with the A-name contributed very little. Because of this, the A-name coauthor is unlikely to agree to randomization, even if she doesn’t care in which order the authors are listed in the absence of negative inference. But with the proposed signal, the certification, I imagine that the random order idea could take off. It could work.

The idea isn’t without its downsides, though. While random order might make it easier to attract collaborators with last names after mine, it might also discourage collaboration by those with last names earlier in the alphabet. An Anderson might be less willing to collaborate, as random order decreases the Anderson’s expected visibility on collaborative projects.  Ultimately, however, I think any such downsides are likely small, and the fairness of the process would certainly be improved.

Overall, I believe this is a really great idea. I’d be happy to do it on any of my future projects. Particularly those papers I’m writing with Agranov and Boleslavsky. But I’d even be happy to do this on the projects where my name would traditionally be listed first.