Criteria for Authorship in the IRRL

Updated 15 July 2010

Contact the Director of the Laboratory if you have questions
Ken Livingston:  livingst@vassar.edu

Projects in science usually involve collaboration.  If the project results in a presentation or published paper there are inevitably questions about who deserves authorship and in what order the authors’ names should be listed.  In order to make decisions about authorship as transparent as possible we have developed a set of principles that are intended to capture relative contributions to a project.  Although we have developed a point system for trying to track these contributions, this system does not eliminate the need to make judgments about the nature and extent of the work done by members of a research team.  At the heart of those judgments is the question of whether a person made a creative contribution to the project, in the broadest sense of the word, “creative.”  That is, did the person actually help to create the final product, which is usually a peer-reviewed paper or presentation.  If so, that person’s contribution is weighted even more heavily if he or she brought special expertise or inventiveness to the process, as indicated by how easily that work could have been done just as well by someone else. 

We divide contributions into six categories corresponding to different phases of the project.  Each phase is allocated a number of points, and the points for all phases sum to 1000.  The points for each phase may be divided among multiple people according to their proportionate contributions.  If someone contributes more than 0 but less than 100 points then he or she will not earn co-authorship but will, instead, be acknowledged in the paper’s footnote.  Anyone contributing 100 points or more will be an author on the paper, with the order of authorship determined by the relative number of points. 

The following are the default point totals for different phases of a project.  However, the project leader may modify these proportions to better fit the unique features of a given project.  These modifications will be published on the lab’s web site in association with the project so that everyone involved knows how the system works. 

The idea phase (150 points):  Every project begins with an idea, a question, or a hypothesis from which the rest of the project is launched.  Note that the foundational idea has to be developed sufficiently to allow the subsequent phases to proceed.  Loose hunches or vague “what ifs” may be the beginning of an idea, but they are not sufficient to complete this phase.  If the idea develops during a discussion, then all who contributed to the final, fully developed idea get credit in proportion to their contributions.

The design phase (100 points):  Details of the design include specification of the conditions under which data will be collected and features of experimental design if true experiments are to be done (control conditions, randomization protocols, counterbalancing, etc).  Poor choices at this phase can be costly and may require duplicated effort and lost time, so this is an important step.

The implementation phase (150 points):  In robotics projects, whether they involve teleoperation, simulation, or autonomous robots, considerable work is always required to assemble the hardware, software and other necessary materials into a working system.  Often the people doing this work are closely supervised so some of the credit may go to the supervisor.

Data collection phase (150 points):  The data collection and implementation phases are often more interleaved than this distinction suggests.  One builds an implementation, tries it out, discovers and documents problems, and tries again.  Nevertheless, these are distinct phases, and different people may be involved in them.  Merely following instructions and repeating a procedure over and over may not earn much credit, but a person who pays attention to what is going on during the process, notices interesting patterns or, especially, problems, and makes constructive contributions to revising the implementation or procedures makes a critical contribution to this phase of the project.

Data analysis phase (100 points):  Just crunching numbers through a spreadsheet or stats package will not earn enough points for authorship, but coming up with a new way to reduce or analyze data, or devising a novel way to present the data are important contributions.  Especially valued are noticing unexpected patterns in the data or having creative insight into better ways to present the data.  Depending on the project and on how routine the data analysis is the maximum points may or may not be allocated.

Writing (300 points):  At the end of the day all of the other work of the project remains a local story with little impact on the field unless someone finds a venue for presenting it and pulls the whole package together into a document or presentation that is accepted for publication.  Writing may be, and often is, the primary responsibility of a single person who has the grand conceptual vision of how the work should be assembled for maximum impact.  A comprehensive and insightful literature review is also an extremely important part of this process.  Note that it is the contribution to the final product that counts here.  Someone might draft a version that is scrapped, for example, and receive no credit as a result, in spite of the amount of time and effort expended.  Results — not effort —  are what count  in this as well as in all phases of the work. 

Final comments:  Cognitive science in general and the field of robotics in particular are multidisciplinary in nature, and different disciplines have different traditions about assigning order of authorship.  For example, in behavioral areas of research the disciplines have published rules for assigning authorship that insist on a strict ordering according to contributions to the project.  Earlier names have always made greater contributions.  In biology, however, as well as in some other cognitive science constituencies, other guidelines may be in play.  For example, if three or more authors are involved, often the person with the second-most points will choose the last position, a place that is sometimes called the “senior” author or “lab head.”  This practice is also common in many European labs.  We are committed to doing our best to allocate authorship fairly and believe that it is easiest to do that if everyone knows what the criteria are from the beginning of the project.  Each contributor to the project is invited to discuss his or her contributions to the project at any phase of the process in order to avoid later misunderstandings.  Differences of opinion about how to evaluate contributions may well develop, of course, and the leader of the project will, in the end, have to make the final judgments about these matters, but the members of the lab are committed to the most equitable process humanly possible.