Opinion
Computing Profession Viewpoint

Teamwork in Computing Research

Considering the benefits and downsides of collaborative research.
Posted
Teamwork in Computing Research, illustration
  1. Introduction
  2. The Case For and Against Teamwork
  3. Doing Teamwork Right
  4. Recommendations
  5. References
  6. Author
  7. Footnotes
  8. Tables
Teamwork in Computing Research, illustration

Teamwork has long been a part of computing research, but now advanced technologies and widespread proficiency with collaboration technologies are creating new opportunities. The capacity to share data, computing resources, and research instruments has been growing steadily, just as predicted when Bill Wulf coined the term “collaboratories” a quarter of a century ago.2,15,16

Teamwork has become the overwhelmingly dominant form of research, so rewarding effective teams, teaching our students how to collaborate, and supporting research on what works and what doesn’t have taken on new importance.

Back to Top

The Case For and Against Teamwork

In recent years, the tools for locating relevant documents, finding team members with special skills, coordinating schedules, and refining reports collaboratively has grown substantially.7 In addition to these technology advances, the willingness and fluency with which young researchers appear to use video and audio conferencing, curated datasets, shared document editors, task managers, and other collaboration tools grows steadily.

Another driving force for teamwork and larger group collaborations is what I see as the increased ambition of team members and the growing expectations from research leaders. Teamwork brings more than larger capacity for work; it opens new possibilities when different disciplines, research methods, and personalities are fruitfully combined.6 In short, research teams of two to 10 members and larger groups of hundreds of researchers, can accomplish much more and conduct higher quality work than they could just two decades ago when the Internet was a novelty.10

The growth of research teams was well documented in a series of papers by Wuchty, Jones, and Uzzi.5,13,14 They reported that from the 1950s to 2000 the average number of journal paper co-authors in science and engineering grew from 1.9 to 3.5. They also showed that the impact as measured by journal citation counts increased as the number of authors grew. Furthermore, the benefit of teamwork grew over time. In 1955, team papers attracted 1.7 times as many citations as solo-authored papers, but by 2000 the advantage grew to 2.1 times as many citations, suggesting that technologies and teamwork skills had enabled teams to be more effective than they were in the past.

A study of the papers for the ACM SIG-KDD 2014 Conference, by program chairs Jure Leskovec and Wei Wang, added evidence of the benefits of teamwork. The reviewer ratings of the 1,036 submitted papers increased steadily for papers with up to five co-authors, then remained level. Reviewer ratings may be imperfect, but this bit of evidence seems potent, especially since this conference had an impressively rigorous acceptance rate of approximately 14%. Another outcome was tied to the ratings for papers. Those papers whose authors included a mix of academics and business practitioners had statistically significantly higher ratings than papers whose authors were either all academics or all business practitioners. This added to the evidence that diversity in teams is a catalyst for high quality.8

However, teamwork has downsides, requiring extra coordination among collaborators, learning new disciplines, adjusting to fresh research methods, and accommodating different personalities.4 These downsides mean that those who engage in teamwork will need to learn how to do it effectively, so as to attain the full benefits.

Back to Top

Doing Teamwork Right

The interest in research teams and larger groups has now accelerated with the publication of the National Academies report Enhancing the Effectiveness of Team Science.3 This report adds recent evidence that teams and larger groups are a growing phenomenon in science and engineering research, where multiple authorship has risen to 90% of all papers in 2013.

The report makes another useful contribution by characterizing seven dimensions that challenge today’s research teams (see the accompanying table). Teams on the left side of the range are easier to manage, while teams on the right side of the range are more difficult to manage, suggesting that these deserve more study.3 The report offers suggestions of how to improve team processes and calls for increased research on teams and larger groups.

Teams and larger groups of academics and practitioners seem likely to be more effective in choosing meaningful problems, forming successful research plans, and in testing hypotheses in living laboratories at scale. Teaming between applied and basic researchers is likely to be a growing trend, as indicated by a recent National Science Foundation program announcement, Algorithms in the Field:11 “Algorithms in the Field encourages closer collaboration between two groups of researchers: (i) theoretical computer science researchers, who focus on the design and analysis of provably efficient and provably accurate algorithms for various computational models; and (ii) applied researchers including a combination of systems and domain experts.”

Teamwork between academics and practitioners can have strong benefits, as does multidisciplinary collaboration within academic communities. A clear testimonial for joint research bridging computer science and other disciplines comes from David Patterson’s review of his 35-plus years running research labs on computer systems: “The psychological support of others also increases the collective courage of a group. Multidisciplinary teams, which increasingly involve disciplines outside computer science, have greater opportunity if they are willing to take chances that individuals and companies will not.”9

Patterson concludes with his vision of the growth of multidisciplinary teams: “Whereas early computing problems were more likely to be solved by a single investigator within a single discipline, I believe the fraction of computing problems requiring multidisciplinary teams will increase.”

There is always room for solitary researchers who wish to pursue their own projects. There are substantial advantages to working alone, but those who learn team skills are more likely to become part of breakthroughs than those who go it alone.

A near-term impediment to teamwork is the difficulty that researchers expect to have when facing hiring, tenure, and promotion committees, who are perceived as having trouble in assessing individuals who contribute to teams. Even team members who have participated in many award-winning papers fear they will find it difficult to convince review committees of their contributions. Being a first author helps gain recognition, as does documenting the role of each team member in the acknowledgments section. Writing a single-author paper, when this is warranted, may also help in many disciplines.

Back to Top

Recommendations

In light of the growing interest in teamwork, appointment, promotion, and tenure committees would do well to update their methods for documenting and assessing teamwork, so as to encourage and reward effective team participation and leadership. One example to follow is the University of Southern California, which has developed guidelines to emphasize a variety of forms of collaborative scholarship and to introduce attribution standards for contributions to larger projects.12

A second step, for computing educators, would be to increase teamwork training and experiences, so as to raise the quality of students’ work. Team projects in undergraduate and graduate courses would train students in using collaboration tools and nurture their communication skills for future professional or research jobs. The National Academies report makes many recommendations that need to be tailored to fit local computing cultures and the seven dimensions on which today’s research teams differ in complexity.

A third step, for government agencies, would be to increase funding to study computing research teams so as to enable leaders to form and manage successful teams. Teamwork is difficult since it requires different skills than working alone, but the potential for greater impact makes teamwork attractive. A strong research agenda would include applied and basic components to understand which incentives and rewards best amplify success within the seven dimensions of the National Academies report.

Back to Top

Back to Top

Back to Top

Back to Top

Tables

UT1 Table. Seven dimensions that challenge today’s research teams.

Back to top

    1. ACM 2014 Conference on Knowledge Discovery and Data Mining (ACM-KDD); www.kdd2014.org

    2. Cerf, V.G. et al. National Collaboratories: Applying Information Technologies for Scientific Research. National Academy Press, Washington, D.C. (1993).

    3. Cooke, N.J. and Hilton, M.L., Eds. Enhancing the Effectiveness of Team Science. National Academies Press, Washington, D.C. (2015); http://www.nap.edu/19007

    4. Cummings, J. and Kiesler, S. Collaborative research across disciplinary and organizational boundaries. Social Studies of Science 35, 5 (2005), 703–722.

    5. Jones, B.F., Wuchty, S., and Uzzi, B. Multi-University research teams: Shifting impact, geography, and stratification in science. Science 322, 5905 (2008), 1259–1262; DOI:10.1126/science.1158357

    6. Nelson, B. The data on diversity. Commun. ACM 57, 11 (Nov. 2014), 86–95.

    7. Olson, J.S. and Olson, G.M. Working Together Apart: Collaboration over the Internet. Morgan & Claypool Publishers, 2013.

    8. Page, S.E. The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies. Princeton University Press, NJ, 2007.

    9. Patterson, D. How to build a bad research center. Commun. ACM 57, 3 (Mar. 2014), 33–36.

    10. Spector, A., Norvig, P., and Petrov, S. Google's hybrid approach to research, Commun. ACM 55, 7 (July 2012), 34–37.

    11. U.S. National Science Foundation Program on Algorithms in the Field (AitF); http://www.nsf.gov/pubs/2015/nsf15515/nsf15515.htm

    12. University of Southern California, Guidelines for Assigning Authorship and for Attributing Contributions to Research Products and Creative Works (Sept. 16, 2011); http://www.usc.edu/academe/acsen/Documents/senate%20news/URC_on_Authorship_and_Attribution_20110916.pdf

    13. Wuchty, S., Jones, B. F., and Uzzi, B., The increasing dominance of teams in production of knowledge. Science 316, 5827 (2007a), 1036–1039. DOI:10.1126/science.1136099

    14. Wuchty, S., Jones, B. F., and Uzzi, B., Why do team-authored papers get cited more? Response. Science 317, 5844, (2007b), 1497–1498.

    15. Wulf, W.A. The collaboratory opportunity. Science 261, 5123 (1993), 854–855; DOI: 10.1126/science.8346438

    16. Wulf, W.A. The National Collaboratory—A White Paper. In Towards a National Collaboratory, the unpublished report of a workshop held at Rockefeller University, March 17–18, 1989 (Joshua Lederberg and Keith Uncapher, co-chairs).

    Thanks to the anonymous reviewers and readers of early drafts, including Nancy Cooke, Gerhard Fischer, Kara Hall, Margaret Hilton, Judy Olson, Scott Page, Dave Patterson, and Jennifer Preece.

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More