Pushing Interdisciplinary in the Study of Groups and Teams

Pushing Interdisciplinarity

The articles in this special issue reflect both the commonalities and distinctions participants (see Table 1). felt and expressed in the Lorentz, Netherlands workshop, “New Frontiers in Analyzing Dynamic Group Interactions: Bridging Social and Computer Science.” ¹ With only a few exceptions, the computer science participants knew (or knew of) the other computer scientists (also known as geeks, a label we used at the workshop), and the group scholar participants (also known as groupies, another label we used at the workshop) knew (or knew of) the other groupies. So, as group research on group formation would suggest, the first few activities were spent in introducing ourselves and our research interests to one another (see Lehmann-Willenbrock, Hung, & Keyton, 2017). While we found some common ground, we also found differences in theory (see Reiter-Palmon, Sinha, Gevers, Odobez, & Volpe, 2017), and workflow (see Allen et al., 2017). This final article of the special issue seeks to set a roadmap for how groupies and geeks can work together.

OPEN IN VIEWER

Table 1.

Lorentz Workshop Participants.

Joseph A. Allen, Department of Psychology, University of Nebraska at Omaha, josephallen@unomaha.edu Elisabeth Andre, Department of Human Centered Multimedia, University of Augsburg, andre@informatik.uni-augsburg.de Stephenson Beck, Department of Communication, North Dakota State University, stephenson.beck@ndsu.edu Piet van den Bossche, Department of Educational Research and Development, Maastricht University, piet.vandenbossche@maastrichtuniversity.nl Claudia Buengeler, Leadership & Management, Amsterdam Business School, c.buengeler@uva.nl Mohamed Chetouani, Department of Signal Processing and Machine Learning for Human-Machine Interactions, Université Pierre et Marie Curie-Paris, mohamed.chetouani@upmc.fr Ming Ming Chiu, Department of Educational Studies, Purdue University, mingmingchiu@gmail.com Colin Fisher, School of Management, University College London, colin.fisher@ucl.ac.uk Josette Gevers, Department of Human Performance Management, Eindhoven University of Technology, J.m.p.gevers@tue.nl Hatice Gunes, Computer Laboratory, University of Cambridge, Hatice.Gunes@cl.cam.ac.uk Dirk Heylen, Human Media Interaction Group, University of Twente, d.k.j.heylen@utwente.nl Hayley Hung, Pattern Recognition & Bioinformatics Group, Delft University of Technology, H.hung@tudelft.nl Joann Keyton, Department of Communication, North Carolina State University, jkeyton@ncsu.edu Florian Klonek, Department of Industrial/Organizational and Social Psychology, Technische Universität Braunschweig, f.klonek@tu-bs.de Jeremy Lee, Department of Electrical Engineering, National Tsing HuaUniversity, cclee@ee.nthu.edu.tw Nale Lehmann-Willenbrock, Department of Experimental and Applied Psychology, Vrije Universiteit Amsterdam, n.lehmann-willenbrock@vu.nl Yoichi Matsuyama, School of Computer Science, Carnegie Mellon University, yoichim@cs.cmu.edu Marc Mehu, Department of Psychology, Webster University Vienna, marc.mehu@webster.ac.at Annika Meinecke, Industrial/Organizational and Social Psychology, Technische Universität Braunschweig, annika.meinecke@tu-bs.de Louis-Philippe Morency, School of Computer Science, Carnegie Mellon University, morency@cs.cmu.edu Jean-Marc Odobez, Perception and Activity Understanding Group, IDIAP Research Institute, odobez@idiap.ch Roni Reiter-Palmon, Department of Psychology, University of Nebraska at Omaha, rreiter-palmon@unomaha.edu Ronald Poppe, Information and Computing Sciences, Utrecht University, r.w.poppe@uu.nl Tanmay Sinha, School of Computer Science, Carnegie Mellon University, tanmays@andrew.cmu.edu Alessandro Vinciarelli, School of Computing Science, University of Glasgow, vincia@dcs.gla.ac.uk Gualtiero Volpe, InfoMus Lab, University of Genova, gualtiero.volpe@unige.it

What Is a Group?

Answering this simple question, What is a group?, sets the foundation, and often, the direction of research on groups. Most social science group scholars will use a definition with the following five characteristics. First, the minimum number of group members is three, as the interaction among three or more team members has considerably greater complexity (including the development of subgroups) than that of a dyad (see Moreland, 2010). Group scholars seldom put an upper limit on team membership as that parameter is typically constrained by the goal, task, or context of the group. Second, group members must be interdependent. That is, group and individual outcomes are influenced by what other group members do. Group members rely on and cooperate with one another to complete the group activity or goal, as the activity would be difficult or impossible for one individual to achieve.

Third, group members must be interacting toward some task or relational goal; however, this goal is not necessarily a formal one. Goals can be short or long term, simple or difficult. Regardless, group goals give group members direction in their interaction and activity. Fourth, group members have some sense of identity as being members of this group; identity may be formalized or informal. Appointing members to a group is not sufficient to manifest group identity; that is, identity is developed among group members, not assigned by those outside the group. Fifth, groups develop some type of role-supported structure that moves the group toward its goal-oriented activities. Structure may be formal or informal, and tends to develop as group members are interacting. One part of group structure is the development of group rules and group norms. For more formal groups, members may be assigned or take on roles (e.g., chairperson, recorder). For less formal groups, members tend to identify with and develop roles that are in line with their interest and talents. In summary, a group is three or more people who work together interdependently on an agreed-upon activity or goal. They identify themselves as members of the group, and they develop structure and roles, based on norms and rules, as they interact and work toward their goal.

Another important aspect of a group or team is the context or environment in which the team interacts. The size and task of a group can be directed by some outside force, influence, or set of rules. Or, group members create the group and add or subtract members until the group is moving effectively toward its goals. Many groups have permeable and fluid boundaries. Groups may be formally or informally connected to other groups—sometimes sequentially in the completion of a task; other times a task is so complex that subgroups take on goal activity that is interdependent. Finally, especially in organizational structures, groups are embedded in a larger task and social culture that dictates, to some degree, its activities.

A third set of considerations is the time and space of group interaction. Groups may have considerable histories and the expectation of a long future. Other groups may meet once to complete their task. But, even for one-time groups (or ad hoc groups), time is present in the interactions of group members as they move from the start to finish of their time together. Too infrequently, groups and group member interaction is analyzed for where and through what media the group does it work. Now groups meet physically and virtually, face-to-face and through text and other mediated channels. Where a group meets and how a group meets is also important to the identity and interactions among group members. Groups in the same context or organization (i.e., a hospital) can be very different (e.g., surgical teams, patient service units) with respect to the defining characteristics of groups. Thus, the environmental aspects of group interactions must also be considered. These differences, of course, have implications for how group members interact with one another. Some groups or teams will use only one medium whereas others communicate face-to-face and over multiple mediated channels.

The points in these last few paragraphs address the importance of context to the study of groups and their interactions. We suggest that groupies make clearer the advantages, intricacies, and difficulties of researching groups in context and that geeks consider the contextual differences in which groups interact when designing and testing new instrumentation and collecting data. The ideal collaboration would join the sophisticated technology of the geeks with the contextual realism of the groupies.

What Literature? Whose Literature?

An obvious difference between the two sets of scholars is their dependence upon and contributions to different sets of literature.

Collaboration Opportunities

Interdisciplinary collaboration is not easy, but it is necessary (see Beck, Meinecke, Matsuyama, & Lee, 2017). There are several ways social scientists can help computer scientists. First, invite computer scientists to develop and present workshops at (mostly) social science conferences. For example, INGRoup (Interdisciplinary Network for Research) is an annual conference, and social scientists are keen to learn and deploy new methods of data collection and analysis. Preconference workshops at INGRoup and other social science disciplines conferences could (a) help to establish relationships between social scientists and computer scientists workshop presenters, as well as (b) introduce new data analytic strategies to social scientists. Likewise, editors of journals that publish group research (e.g., Small Group Research, Group Dynamics) could invite computer scientists to develop and submit high-level tutorials on new methods for analyzing data commonly used by group scholars (e.g., audio and video recordings, transcripts, weblogs). These are important steps as computer scientists need greater working knowledge with why and how group scholars collect the data they do.

Groupies could help computer scientists by better articulating the relationships between constructs and operationalizations of the variables they study. Developing a standardized method of notating group data would likely yield productive results for both groups of scholars.

When group scholars have the opportunity to collect a significant number of cases of single modal or multimodal data, they should invite a computer scientist to be on the team. Doing so would increase the likelihood that corpora of data could be developed, and used in the future for secondary analyses by both groups of scholars (see, for example, the Collaborative Interaction Corpus; https://sites.google.com/ncsu.edu/cic/home; and the AMI Corpus; http://groups.inf.ed.ac.uk/ami/corpus/). It may be that social scientists are leaving something on the table during their data collections. With a bit of advice and friendly intervention by computer scientists, it is likely more data could be digitized for further examination, and perhaps at a different scale.

When geeks are designing data collection, they could benefit from interacting with groupies. As one example, tracking head movements give groupies a new source of data. That data collection would be more sophisticated if head tracking also included head tilts and leans which are likely indicators of moods and emotions of group members. Finally, just having an overview of a participant’s image or likeness is not enough. Are group members looking down? Are they asleep? Or winking? These details will influence subsequent group interactions.

Buengeler, Klonek, Lehmann-Willenbrock, Morency, and Poppe (2017) offer several ideas for killer apps that would require group and geek collaboration. Not only would these apps enhance interdisciplinary collaboration, they would serve as prominent public-facing examples of the value of interdisciplinary collaboration between geeks and groupies.

Amending an idea from the ACM International Conference on Supporting Group Work playbook (see Design Fictions—Fictive Futures: Exploring Future Research Agendas; http://group.acm.org/conferences/group18/CFP.pdf), conference planners and journal editors could seek submissions that address future collaboration of geeks and groupies. These interdisciplinary proposals—written by geeks and groupies, and based on both social science and computer science literature—would propose a data collection and analysis that is not presently possible. Paneling such proposals as brainstorming sessions has the potential of finding solutions and creating new interdisciplinary relationships.

Make Interdisciplinary Collaboration a Reality

There are many significant outcomes to be achieved by these suggested activities. The first is to enable more sophisticated analyses of data collected. Group scholars need additional technological tools for more sophisticated analyses of their rich data; computer scientists need stronger theory for what should be analyzed with technology. Working together, both groups of scholars will benefit from strengthening the link between data analysis and theory building.

But to make that happen, these communities of scholars must have regular and goal-oriented conversations with other another. Three candidate joint projects are (a) developing corpora of data for use by both groupies and geeks, (b) developing common theoretical frameworks to ensure that digital data collection addresses or tests theoretical assertions at the group level (see Reiter-Palmon et al., 2017), and (c) integrating the workflow of groupies and geeks. Indeed, this last point may be the most difficult. We have addressed it in this special issue (see Allen et al., 2017), and others have also advocated the integration of workflows (e.g., Singh & Vouk, n.d.).

Computer scientists and group scholars are not the first to propose research collaboration (see Fiore, 2008; Salazar, Lant, Fiore, & Salas, 2012). A group of social scientists were invited to create the first Science of Team Science conference (see http://www.scienceofteamscience.org/). This conference has continued, a large body of team science literature has accumulated, and procedures and tools to help interdisciplinary teams perform more effectively is available to the public (https://www.teamsciencetoolkit.cancer.gov). The founding team members took the initiative to identify a roadmap for team science research that also scaled the relative importance of the research areas on the roadmap (see Falk-Krzesinski et al., 2011).

Within the last few years, health care practitioners and scholars have tried a slightly different approach to decrease infection, complication, and death rates connected with cancer. With the power and influence of the U.S. National Cancer Institute and the America Society of Clinical Oncologists (https://healthcaredelivery.cancer.gov/healthcare/nci-asco.html), teams of oncologists, oncology nurses, and patient advocates were paired with team and group scholars for working sessions. Each team created a case study that incorporated oncology health care issues and the group issues that evolve from these multispecialty cases. Several case studies were selected for publication in the November 2016 issue of Journal of Oncology Practice (see Johnson, Macpherson, Smith, Block, & Keyton, 2016); others became the focus of online workshops for health practitioners.

The instrumental first steps, we believe, are to have computer scientists and group scholars collaboratively map a research agenda and develop proposals for funding for the ideas that will strengthen the connection between these two scholarly communities.