Over the past two years, I have been leading a collaboration of 4
universities. We are now writing a bunch of papers and giving talks, so
naturally, my graduate students are asking me who should be included as
coauthors?. Almost three decades ago, this same question came up while I
worked at Bell Labs as part of a large team. Since there was pressure to publish, conflict was inevitable when a researcher was dropped from the author list.
One of my
collaborators found a document that spelled out reasonable guidelines. I have
not been able to locate my copy -- since it was on paper, I probably lost it -- but here is what I remember.
Collaboration is integral to science, reaping rewards that exceed the
sum of individual efforts. Success thus requires an understanding
between collaborators that researchers who do the work can expect to be
coauthors. There are many types of contributions, and all are critical,
so it is pointless to argue which are more important. To this end,
the criteria for coauthorship can be parsed into five broad classes.
Coauthorship requires that the individual contribute substantively to at
least three of these five. The five criteria, in typical chronological
order follow:
1. Having the idea. All research is initiated by a question and a
rough outline of activities that will lead to an answer. For example,
recent data show evidence that the earth is warming. A stroke of
insight leads to the realization that a temperature record is literally
frozen in place in polar ice caps that goes way back in time. The
scientist then imagines ways that a core can be extracted and how they will be evaluated
to determine a historic temperature profile.
2. Getting support for the work. Research requires resources. An
individual like Albert Einstein had deep insights and worked out his
ideas at home after work as a patent examiner. In this case, he
supported his own research. Collecting ice core samples, on the other
hand, requires expensive expeditions, specialized
equipment, and technical expertise. A scientist proposing such
experiments to a funding agency must succinctly pose the question to be
answered, provide a meticulous plan for experimental protocols with
well-defined timelines, and estimate the cost of the project, including
manpower needs. Getting funding is a time-consuming activity with low chances of success, so is an important ingredient to research.
3. Doing the research. Once the idea is generated and funding secured
for the project, someone has to do the work. To get credit for doing
the research, one must do a substantial chunk of the work and should also
contribute both intellect and muscle. If a technician is given specific
instructions, such as turn this knob and write down this number, that
doesn't count even if (s)he spends a significant amount of time and
effort. Acting like a machine taking data and being a computer
following code is not the creative activity we call research. On the other hand, interacting with a colleague or advisor who guides the direction of the work warrants co-authorship even if that individual did not work in the lab, analyze data or wrote code.
4. Writing the paper. All researchers who have made a contribution to a
piece of work will typically provide input to the manuscript. Writing
the paper implies a significant amount of work that includes
interpreting the results and tying it all together. Proofing the
manuscript is required of coauthors but correcting typos is not sufficient to count as
writing the paper. Adding an overall wrapping to the work that ties it together is part of the creative process.
5. Understanding the paper. This is less of an activity and more the
result of being involved in the day-to-day work in generating results
and writing the paper. Each coauthor should have a good enough grasp of
the science to be able to explain the work to other scientists. There
are always exceptions, especially when doing interdisciplinary work. A
chemist may need to muster loads of creative juices to figure out how
to make a compound that a Physicist measures to get the result. It is
unreasonable to require the physicist to understand synthetic chemistry
and the chemist should not be required to understand relativistic
quantum mechanics. If the chemist writes a strictly synthesis paper,
the physicist need not be a coauthor unless the reason for making the
material is integral to the argument for making the material.
Similarly, the chemist need not be included as a coauthor in a series of
physics papers that use the chemist's material. However, if the
chemist makes a material with the specific requirements for the physics
experiment, that counts as creative input, and the chemist should be
included as a coauthor provided he or she has contributed to two other categories.
These guidelines are not meant to be applied as absolute rules. Nor
should they be used as an excuse to cut off a potential coauthor. Rather,
they should be the starting point for discussions. There are
undoubtedly many exceptions for a diversity of reasons, so treat each of these thoughtfully. When in doubt, err on the side of including
your collaborators as coauthors. It never hurts to add a coauthor to a
paper. However, it is unethical to add undeserving authors.
There are many reasons for not adding coauthors. Years ago it was
common practice to add the name of a highly-respected colleague to paper
to give it added credibility. As a result, high-profile publications
now ask who did what in the reported research. Clearly, it is unethical
to make deals with collaborators to add my name to your papers in expectations of
reciprocation.
Then there are the fuzzy areas. For example, it often happens that a
researcher worked on a project side-by-side with team mates, but his or
her part didn't pan out. If this person is involved in writing the
paper and has a firm grasp of the results, then he or she should be a
coauthor.
I am sure that you can think of many exceptions and gray areas. Treat
them thoughtfully and you will have good collaborators for a lifetime. Let me know what you think.
