Who Gets Credit?

In the physical and biological sciences, it's common for papers in journals to have multiple authors -- sometimes dozens of them -- and departments have long accepted that C.V.'s will be full of jointly produced work. In many other fields, work has traditionally been more solitary. Look at this year's issues of the American Historical Review, for example, and not a single article or review essay has more than one author.

Political science historically has been a field more like history, with single-author work the norm. But increasingly, political scientists are writing together -- and that has led the American Political Science Association to start a discussion on the implications this has for the faculty members and graduate students involved.

The association wants to talk about such issues as whose name goes first in a paper -- a question that might seem minor, but may not be to a candidate up for a job or for tenure. More broadly, the association wants professors to talk about how collaboration is taught to graduate students. A physicist or biologist can only go so far before being part of a lab team -- should the same be true of a political scientist?

The American Political Science Association appointed a special panel to consider these and other issues, and its report has just been released. The report documents the shifts in political science, tries to summarize the issues that these shifts raise, and offers some suggestions on policy areas. The association will sponsor a special discussion of these issues at its annual meeting later this summer as well.

"What we are trying to do is to document the patterns and think through the ethics of these issues," said Kanchan Chandra, a political scientist at New York University who led the panel.

Befitting a discipline that studies power, one of the key issues raised by the project so far is that much of the collaboration is "asymmetrical," meaning that its involves a tenured and a non-tenured professor, or a professor and graduate student. Generally, the panel's report suggests issues for discussion rather than seeking to specify certain policies as appropriate.

But the importance of the issue of unequal partnerships to the panel is evident in that it was one of the few places where it made a specific recommendation: The panel says that given the awkwardness of discussions about who gets credit for what, junior partners should not have to be the ones to raise the issue, and that it should be considered the responsibility of a senior partner to do so.

Political scientists are not the only discipline to think about the impact of collaboration -- although fields include some where discussions are far less developed and others where issues are largely taken for granted. A report on tenure policies issued last year by the Modern Language Association, noted that "solitary scholarship, the paradigm of one-author-one-work, is deeply embedded in the practices of humanities scholarship," but questioned whether that paradigm is always appropriate. The MLA panel noted that digital scholarship has led more professors to work together and called on departments evaluating candidates for tenure and promotion to focus on the quality of work. Jointly produced work, the report said, "should be welcomed rather than treated with suspicion because of traditional prejudices or the difficulty of assigning credit."

If collaborative work is still new for some disciplines, it is standard elsewhere and protocols are generally understood, even if they aren't codified. Of the major articles in the latest issue of American Economic Review, six are by single authors, seven by two authors, two by three authors, and one by four authors. All of the multiple author pieces list names alphabetically.

Robert Moffitt, editor of the journal and a professor of economics at Johns Hopkins University, said that journal editors in economics almost always leave such questions to authors to decide themselves and that there is "a strong social norm" to list names alphabetically. There are "occasional deviations," he said, "where the relative contributions of the authors is particularly disproportionate," and he estimated that in his career, maybe 3-5 percent of the articles on which he was a co-author didn't list names in alphabetical order.

Part of the motivation for political science taking up these questions is that the shifts in that field -- from solo being the norm to joint papers becoming common -- have happened gradually over decades, and aren't the same in all parts of the discipline. As a result, there is less of the social norm than in economics.

The panel that studied the issue analyzed journal articles across political subfields, and found that while less than 10 percent of articles had multiple authors in the decade of 1956-65, about 40 percent did in 1996-2005. Combining fields, however, may understate the relatively recent change in key subfields. Journals in political theory have never embraced collaborative work and only about 5 percent of articles have more than one author. But in the last decade, the report notes, co-authorship has become the norm, and covers a majority of articles in top journals in American politics.

Another change the panel noted is the proliferation of "team" research projects. The concept of such projects isn't new and some have been around for decades, the panel said, citing such examples as American National Election Studies, based at the University of Michigan. But the APSA panel said that there are many more large-scale research programs now, citing as an examples work at Columbia University on the initiation and termination of war.

On the issue of who collaborates, the panel analyzed the papers presented at the association's annual meeting and found that most do not involve academics on equal footing.

Collaborations on APSA Meeting Papers, 2006

After documenting that collaboration has arrived in political science, the association's panel identified five key questions that it thinks merit more consideration:

• How should the contribution of assistants be acknowledged in collaborative work?
• What are the criteria by which an assistant's contribution to a project should be acknowledged as co-authorship?
• What should the order of authors in a co-authored work be?
• How can we integrate collaborative work with graduate training in a way that encourages independent thinking?
• What should the procedures be for a discussion of any of these questions and for the resolution of disputes.

The tone of the report is not prescriptive, but rather aims to suggest topics of conversation. For those who assume that some matters are settled because they have experienced collaboration in a certain way, the panel may have a few surprises. For example, while it found that 69 percent of jointly written articles in American Political Science Review list authors in alphabetical order, it found that in some cases, authors who work together on multiple projects order their names randomly across various articles.

The panel appeared to be especially concerned to raise issues that might be difficult for some junior partners to voice. For example, in documenting practices for acknowledging the work of assistants, it found that some scholars consider that no acknowledgment is necessary as long as the assistant is paid. The panel rejected this approach and noted that the APSA's ethics committee has found that research assistance should be acknowledged -- irrespective of the question of pay.

Looking to the future, the panel also explored how the trend of collaboration plays into graduate education. Graduate departments do not welcome collaborative dissertations, and the panel didn't suggest that they start doing so. With the dissertation off limits for collaboration, many graduate students end up collaborating with their advisers on other work -- and this can have a range of impacts, the panel found.

While it exposes graduate students to collaboration, it is rarely equal collaboration and it can divert a graduate advisor's time and attention from a student's dissertation. One graduate student told the panel: "My concern about the collaborative relationship with my adviser was not about the allocation of credit but about the use of my time. Once we began working together, it was as if my dissertation did not exist. My advisor simply had no mental space to discuss it."

The panel urged graduate programs to experiment with different approaches to promote collaboration and also to respect the interests of graduate students. One model suggested was to encourage collaboratively collected data for a dissertation, but to insist that the dissertation itself be on an independently generated idea. Another approach would be for professors to use first- and second-year graduate students as "apprentices" on research, but to explicitly phase out that research as the students reach the point where they need to focus on their dissertations. Yet another approach might be to use a "team" model for advising, where professors working together might jointly advise a graduate student. "Collaborations of this nature can encourage independence by exposing a student to a wide range of intellectual influences and allowing students to create a network of professional relationships," the report said.

While the report noted many positive aspects to collaboration, it also acknowledged that on questions of who gets credit for what, disagreements will occur. The political scientists concluded their study with a call to avoid warfare over such matters. "It is almost always better to give up a claim on co-authorship, or to relent and let someone else be a co-author," the report said, "if that is what it comes down to, since no one contribution in political science is going to win a Nobel Prize and the next article can almost always be better."