As an empirical social scientist, the conversation about interdisciplinarity has frustrated me for years. The National Academies of Sciences produced a volume in 2004 titled Facilitating Interdisciplinary Research, in which they delineated ways to support interdisciplinary research (IDR), but neglected to document that such research was, indeed, beneficial to scientists or society—or at least better than disciplinary work.1Read online → They appeared to conflate interdisciplinarity and innovation, though the two are not synonymous (as Jerry Jacobs notes in a previous contribution to this series). This oversight was understandable: at the time, there was precious little research documenting the effects of IDR.
“Specializing helps scholars in almost all regards, except in terms of visibility.”But my own research program was evolving in this direction. As an assistant professor, I was told I was too scattered and that my diverse research program (and I) needed some focused identity if I were to attain tenure. Never one to follow advice without empirical proof, I embarked on a project to measure the extent of specialization in scholars’ research program, and whether and how it influenced career outcomes (n.b., this was also a natural outgrowth of my dissertation). I found that specializing helps scholars in almost all regards, except in terms of visibility; in this realm, generalist scholars with diverse research programs are more advantaged. Because my measure of specialization could not distinguish between two kinds of generalists—those who wrote multiple papers on distinct topics, and those whose papers integrated two or more subfields—I developed ways to measure subfield integration and assessed its influence on scholarly visibility. Coauthor Jim Moody and I found that the mixing of subfields, especially those that are rarely combined, makes a paper more visible in terms of citation counts.2 Leahey, Erin, and James Moody (2014). “Sociological Innovation through Subfield Integration.” Social Currents 1(3): 228-256. We concluded that subfield integration is a means by which innovative, noteworthy research arises.
This research background, combined with the policy discourse about interdisciplinarity, prompted me to ask: How does interdisciplinary research influence scholarship and scholarly careers? I began to brainstorm about research designs that would help me answer this important science policy question, and discussed these in my fellowship application to the Radcliffe Institute for Advanced Study at Harvard. Because of Radcliffe’s commitment to multidisciplinary engagement, I thought for sure my proposal—which expressed concerns about the hype surrounding interdisciplinarity and a desire to empirically ground conversations surrounding IDR—didn’t stand a chance. Happily, I was wrong. Presenting my preliminary ideas a month after my arrival at Radcliffe, in October 2008, I was delighted to receive some of the most astute feedback from the poets, physicists, computer scientists, and marine biologists who formed my diverse fellowship community.
Since then, I’ve devoted much of my research program to investigating just these effects, with the hope of informing discussions of interdisciplinarity, providing guidance to scholars’ career decisions, and grounding science policy. Below, I elaborate on two of these studies, as well as a current NSF-funded project.
The “productivity penalty” and other challenges of practicing IDR
My first project on interdisciplinarity revealed that scholars with greater involvement with IDR are indeed prominent (as indicated by citations), but also less productive. In a paper in press at the Administrative Science Quarterly, Christine Beckman, Taryn Stanko, and I investigated both possible costs and benefits of engaging in IDR.3 Leahey, Erin, Christine Beckman, and Taryn Stanko (in press). “Prominent but Less Productive: The Impact of Interdisciplinarity on Scientists’ Research.” Administrative Science Quarterly.
Read online → One body of literature, organizational ecology, suggests that spanning disciplines is detrimental because interdisciplinary offerings get caught in the middle and can’t be viewed and evaluated clearly. This, we suggest, makes it harder for scholars to navigate the peer-review system and publish IDR. Another body of literature, on innovation, suggests that spanning disciplines is beneficial and a foundation for new ideas. This, we suggest, makes research certified by the peer-review process more interesting and useful to the scientific community. Thus we hypothesized both costs (lower productivity) and benefits (higher visibility) to IDR.
We also pushed this literature forward by moving away from a dichotomous concept of spanning to a continuous one that considers the degree of similarity between spanned fields. Are they cognate fields, like sociology and anthropology? Or are they cognitively distant, like public health and optics? Spanning distant fields may be harder, but the payoff might also be greater.
We tested these ideas on a sample of nine hundred scientists from many different fields. We collected data on them and their combined thirty thousand published articles from CVs, websites, the Web of Science, and other databases. From the bibliography of each article, we calculated Alan Porter’s continuous IDR score that considers not just the variety of fields represented in the bibliography, but also their balance and their degree of dissimilarity.4→ Porter, Alan L., Alex S. Cohen, J. David Roessner, and Marty Perreault. 2007. “Measuring Research Interdisciplinarity.” Scientometrics 72:117-147.
→ Porter, Alan L. and Ismael Rafols. 2009. “Is Science Becoming More Interdisciplinary? Measuring and Mapping Six Research Fields Over Time.” Scientometrics 81:719-745. We then assessed whether and how IDR affects productivity and visibility.
“Indeed there is a trade-off to pursing IDR: it increases visibility but dampens productivity.”
Our analyses revealed that indeed there is a trade-off to pursing IDR: it increases visibility but dampens productivity. We also showed that the relationship between the spanned fields matters, especially for productivity. It is more difficult to produce and successfully publish scholarship that spans unrelated fields, such as chemical engineering and anthropology, than related fields, such as chemical and civil engineering. Spanning unrelated fields improves citations, but spanning related fields does too (if only slightly), presumably by broadening one’s prospective audience.
With supplementary data from various sources, we were able to explore two possible mechanisms behind the documented “productivity penalty.” IDR could stifle productivity because it is penalized in the review process: reviewers and editors may not have the skills and experience to review and appreciate interdisciplinary research. But analyses of data from two sources fail to provide support for this possibility. First, we find no significant correlation between IDR score and days under review, suggesting interdisciplinary papers did not need more help with paper development than more disciplinary papers. Second, a comparison of working papers obtained from arXiv reveals that eventually published working papers are actually more interdisciplinary than working papers that remain unpublished. IDR papers do not appear to be hindered in the review process.
“Scholars conducting multidisciplinary research are more productive, not less.”IDR could also stifle productivity because the process of producing interdisciplinary scholarship—learning new concepts, literatures, and techniques, working with a diverse group of collaborators—is challenging. We find that IDR projects indeed face these hurdles. Scholars who engage in “repeat collaborations” with the same set (or subset) of authors experience a smaller productivity penalty than we see overall, so scholars do become accustomed to working with their coauthors (regardless of field). Survey responses from a small sub-sample of these scholars reveal that interdisciplinary teams have more difficulty generating ideas, and their communication is less clear, more difficult, and lower quality. Last, because our theory suggests that it is challenging to incorporate different ideas in a single paper, multidisciplinary scholars (who publish separate papers on distinct topics) likely do not face production penalties. Indeed, our results show that scholars conducting multidisciplinary research are more productive, not less.
Taken together, our foray into mechanisms reveals that the review stage may not be such a roadblock; we do not find IDR work more likely to be tossed in a file drawer or rejected. Rather, the initial stages of production, during which authors plan, coordinate, and conduct their research, is the largest hurdle for interdisciplinary work, and thus cognitive and communication obstacles are the main source of the production penalty.
In a second project, I explored other possible negative repercussions of IDR more closely in an interview-based study of successful interdisciplinary scholars in the humanities. In a chapter in an edited volume titled Investigating Interdisciplinary Research: Theory and Practice across Disciplines, David McBee and I reveal the ways in which recipients of the Mellon Foundation’s “New Directions” Fellowship struggled to integrate their new direction with their old one.
Reduced productivity was mentioned by almost half of the interviewees, and other more prominent themes help us understand why IDR dampens productivity. Almost 90 percent of interviewees lamented that framing research for a new and different audience is challenging, as is the cognitive work to grasp new areas of scholarship (most of these scholars published sole-authored work). One scholar revealed that “there were quite a few years that my nightmare was that the Jewish historians would dismiss it as the work of a Russianist, and the Russianists would dismiss it as the work of a Jewish historian,” while another admitted that “you feel like an idiot when you expose yourself to being judged by other standards.”
Almost 80 percent agreed that IDR takes longer to produce than monodisciplinary work, for example, “In terms of promotion, [IDR is] not going to harm my prospects, but it will make it slower, you know … it’s going to be a few more years until I’m promoted to full professor.”
Almost two-thirds of the New Directions fellows we interviewed perceived a lack of support from members of their home discipline, who may not understand, appreciate, or support interdisciplinary work. For example, one philosopher whose “new direction” is mathematics described being teased by a fellow philosopher in his department: “…he jokingly said to me, ‘What new thing are you going to learn for the Mellon scholarship? Are you going to learn philosophy?’”
Our results suggest that it is difficult, and perhaps rare, to attain the shared cognitive-emotional-interactional (SCEI) platform that Michèle Lamont and colleagues find ideal. Our study reveals that the challenges that individual scholars face when pursuing IDR (even when supported handsomely by a prestigious fellowship) are varied and go beyond lower productivity.
Understanding universities’ commitment to interdisciplinary research
“How do universities that pursue interdisciplinary research fare?” It is not only IDR’s implications for scholarly careers that matter. How do universities that pursue IDR fare? In a current NSF-funded project with colleague Sondra Barringer, I am trying to answer just this question.5 National Science Foundation SciSIP Program ($270,691), Lead Co-PI on “Collaborative Research: University Commitment to Interdisciplinary Research: Scope, Causes, and Consequences” (award #1461989) with Sondra Barringer. 2015-2018.More info →
We are collecting data on over 150 top universities nationwide, and developing measures of what we call university commitment to IDR from data on the nature of departments, the prevalence of research centers, and from faculty job postings. For example, following Jerry Jacobs and Scott Frickel in their 2009 Annual Review of Sociology paper titled “Interdisciplinarity: A Critical Assessment,” we will calculate the ratio of departments (traditional academic structures) to research centers (more modern and interstitial structures). Following Steve Brint’s work,6 See, for example:
Brint, S. G., L. Turk-Bicakci, K. Proctor, and S. P. Murphy (2009). “Expanding the Social Frame of Knowledge: Interdisciplinary, Degree-Granting Fields in American Colleges and Universities, 1975–2000.” The Review of Higher Education, 32: 155-183. we will elaborate on this by distinguishing more and less interdisciplinary departments, and by not assuming that all research centers are inherently interdisciplinary. Preliminary analyses suggest that there is broad variation in universities’ commitment to interdisciplinary research.
What explains variation in university commitment to IDR? We are pursuing a creative approach to data and analysis to examine the role of precursors—both top-down forces that encourage faculty to pursue interdisciplinary research, as well as bottom-up inclinations from faculty themselves. Top-down incentives are likely apparent in university documents like annual presidential addresses and more episodic strategic plans. We have collected such documents, and using Natural Language Processing techniques like semi-supervised machine learning to classify them in terms of both mentions of and sentiments toward IDR. In addition to policies and plans, faculty members’ pursuit in IDR itself can push the university toward greater engagement. To capture bottom-up drivers of IDR, we are gathering data on faculty publications and external grants from NSF.
“Do the trade-offs we see at the individual level transfer to the organizational level?”We will also examine whether the trade-offs we see at the individual level transfer to the organizational level. Do universities that are highly committed to interdisciplinary research also experience declines in productivity, and an increase in the visibility of scholarly work produced there? We will also assess whether IDR affects other important economic and social outcomes. In terms of economic outcomes, previous research has shown that actors who span domains experience more growth, improved performance, and a competitive advantage, and we will investigate whether income derived from commercialization efforts like patenting and licensing hold for universities that are highly committed to IDR. In terms of social outcomes, previous research finds that domain-spanning actors are rated more favorably, and their output is deemed higher quality. To assess whether these ideas scale up, we will examine whether university commitment to IDR positively influences university prestige ratings.
I feel fortunate that a topic that evolved so organically out of my prior research (and, perhaps, anxiety about tenure), and that is so inherently interesting to me, is also timely. Most of my previous work, I fear, was characterized as boutique sociology with little significance outside the ivory tower. Although Andy Abbott, among others, reminds us that interdisciplinarity is not new (as the SSRC is credited with coining the term in 1926), it certainly has a lot of cachet of late. I am happy to contribute to the development of a body of empirical scientific work on interdisciplinarity and its effects.
References:
Read online →
→ Porter, Alan L. and Ismael Rafols. 2009. “Is Science Becoming More Interdisciplinary? Measuring and Mapping Six Research Fields Over Time.” Scientometrics 81:719-745.
More info →
Brint, S. G., L. Turk-Bicakci, K. Proctor, and S. P. Murphy (2009). “Expanding the Social Frame of Knowledge: Interdisciplinary, Degree-Granting Fields in American Colleges and Universities, 1975–2000.” The Review of Higher Education, 32: 155-183.