2
Science to Inform Team Science
The preceding chapter defined “team science” as scientific collaboration conducted by more than one individual in an interdependent fashion. It also identified seven features that create challenges for team science. This chapter focuses on two of the scientific fields that have centrally contributed diverse methodological and conceptual approaches to understanding and addressing these challenges. Together, these fields provide cumulative empirical knowledge to assist scientists, administrators, funding agencies, and policy makers in improving the effectiveness of team science. We first discuss the social science research on groups and teams and then the “science of team science,” an emerging, interdisciplinary field focusing, as its name suggests, specifically on team science.
This report draws heavily from the social science literature of groups and teams. Organizational, cognitive, and social psychologists have studied team processes and outcomes for more than four decades, providing strong evidence about processes that enhance team performance and how those processes can be influenced (e.g., Kozlowski and Ilgen, 2006; Mathieu et al., 2008; Salas, Cooke, and Gorman, 2010; see also Chapter 3 in this report). As noted in the previous chapter, much of this research focuses on teams in contexts outside of science, yet these teams in other contexts incorporate many of the key features that create challenges for team science. In addition, emerging research focusing specifically on science contexts is beginning to identify similar processes to those identified in other contexts. Thus,
this research is relevant to science teams, and we draw extensively on it in Chapters 3 through 6. In addition, some studies have focused specifically on industrial research and development teams, which are typically composed of scientists engaged in research, similar to academic science teams. For example, Bain, Mann, and Pirola-Merlo (2001) examined the relationship between team climate and performance in research and development teams, and Keller (2006) studied leadership in research and development product teams.
Research on groups and teams has benefitted from the use of simulation and modeling, and it is likely that research on team science can benefit similarly. Simulation allows technological tasks conducted by science teams in the real world (e.g., joint use of scientific equipment or virtual meeting technologies) to be studied under controlled laboratory conditions (Schiflett et al., 2004). For instance, simulation can be used to mock up technologies that human users interact with in the laboratory. One or more technologies can then be evaluated on usability as well as on their ability to improve effectiveness in a science team or group. In addition, agent-based modeling, dynamical systems modeling, social network modeling, and other forms of computational modeling have become more prevalent in the teams literature and can help to extend empirical results from small science teams to larger groups of scientists and scientific organizations (National Research Council, 2008; Gorman, Amazeen, and Cooke, 2010; Kozlowski et al., 2013; Rajivan, Janssen, and Cooke, 2013).
The complex and variegated nature of team science makes the scientific investigation of all its dimensions and contexts quite challenging. Toward the goal of better understanding these inherent complexities, a new field, the science of team science, has emerged (e.g., Croyle, 2008; Stokols et al., 2008a; Fiore, 2008, 2013). In this chapter, we identify some of the unique concerns and contours of this rapidly expanding field, which has been defined as:
a new interdisciplinary field . . . which aims to better understand the circumstances that facilitate or hinder effective team-based research and practice and to identify the unique outcomes of these approaches in the areas of productivity, innovation, and translation. (Stokols et al., 2013, p. 4)
While drawing heavily on the perspectives and findings from research on groups and teams, scholars in the science of team science are concerned with a number of questions that have not been addressed explicitly in that research, as discussed below.
Distinctive Concerns of the Science of Team Science
The scholarly and applied concerns of the science of team science are closely related to the seven features outlined in Chapter 1 that can pose challenges. The distinctive concerns of the field include
- focusing on highly diverse units of analysis, ranging from the level of the team to broader organizational, institutional, and science policy contexts, including centers and institutes specifically designed to promote and sustain team science;
- understanding the multinetwork structure of scientific collaboration, including the diverse contexts and pathways of collaboration that have emerged in recent years;
- understanding the promise and challenges of diverse team membership and deep knowledge integration, especially in transdisciplinary projects that aim to achieve practical as well as scientific innovations;
- establishing reliable, valid consensus criteria for evaluating team science processes and outcomes; and
- focusing on translational and educational as well as scientific goals.
Focusing on Highly Diverse Units of Analysis
Team science encompasses an enormously diverse set of arrangements for conducting collaborative science. As discussed in the previous chapter, team science projects vary in size, duration, level of funding, geographic dispersion, and level of disciplinary integration (Stokols, 2013). Reflecting this diversity, the field focuses on multiple, interacting levels, posing challenges for theory and research.
First, at a team level of analysis, the science of team science field focuses on science teams and groups and their individual members as the principal units of study. Chapter 3 reviews various individual- and team-level factors that influence the functioning and outputs of science teams and larger groups.
As the field’s focus moves beyond individual science teams to higher levels of analysis, it focuses on a variety of organizations and institutions whose mission or goals are to facilitate and sustain effective team science collaboration (Börner et al., 2010; Falk-Krzesinski et al., 2011). For example, universities often establish new research centers focusing on particular scientific and societal problems (e.g., cancer control and prevention; environmental sustainability) to facilitate cross-disciplinary team-based research addressing these problems. Such centers often support several different sci-
ence teams that may work together in pursuit of shared research goals as part of a multiteam system (DeChurch and Zaccaro, 2013).
In addition to its special focus on organizations such as research centers, the science of team science seeks to understand more generally the extent to which various scientific organizations and institutions (e.g., research universities, national laboratories, research funding agencies) may support or hinder team science (see Chapter 8 for further discussion). For example, researchers might analyze how research university incentive structures, such as promotion and tenure policies, affect scientists’ motivation to participate in team science. As another example, one recent study assessed the relative scientific productivity rates of tobacco scientists participating in National Cancer Institute Transdisciplinary Tobacco Use Research Centers (TTURCs) with those of National Institutes of Health grantees working on the same topics as members of smaller research teams who are not participating in the broader research centers (Hall et al., 2012b). Such questions about the effectiveness of alternative research infrastructures or the translational impacts of team science programs have not been explicitly addressed in earlier research on non-science teams.
Finally, at the broadest level of analysis, the field is concerned with how community and societal factors, including social, cultural, political, and economic trends, influence decisions to use a team science approach, the selection of phenomena to be investigated, and the prospects for successful collaboration in the investigation (Institute of Medicine, 2013). For example, policy makers, health care professionals, and scientists are currently focused on ameliorating the national trend of increasing obesity with its attendant adverse health effects (e.g., Institute of Medicine, 2010). Here, science policy concerns rise to the fore, as researchers study the design of funding mechanisms to encourage and sustain science teams and groups, as well as peer review and program evaluation criteria (e.g., Holbrook, 2013; Jordan, 2013) for judging the effectiveness of such teams and groups (see Chapter 9 for further discussion).
Understanding the Multinetwork Structure of Contemporary Scientific Collaboration
Social scientists have begun to investigate the important role of networks in advancing scientific knowledge. For example, sociologist Randall Collins (1998) conducted a comprehensive sociological analysis of the intellectual debates and relationships within and among networks of scholars since the time of the ancient Greeks, arguing that these networks have catalyzed major intellectual advances in philosophy, science, and other fields. In another example, Mullins (1972) traced the creation of molecular biology as a new scientific discipline in the 1960s to the evolving networks of rela-
tionships among a group of colleagues, students, and co-authors studying the bacteriophage, a virus that infects bacteria.
Today, team science increasingly takes place through multiple networks and teams that may be closely linked or unrelated. A given scientist may participate to varying degrees in these networks and teams. The science of team science field is concerned with understanding this multinetwork structure of scientific collaboration in the early 21st century (Shrum, Genuth, and Chompalov, 2007; Dickinson and Bonney, 2012; Nielsen, 2012). Scientists often simultaneously participate in multiple teams, and these teams are embedded within larger networks that are based on their past collaborations (Guimera et al., 2005). These large scientific and translational networks include closely linked groups of individuals who have conducted research and perhaps published together, and also more loosely affiliated groups. For example, some members of the committee that authored this report previously collaborated extensively with others to evaluate National Cancer Institute team science projects (e.g., Stokols, Hall, and Vogel, 2013), others are affiliated through their shared membership in the National Academy of Sciences, and still others are affiliated as faculty members at the same universities.
Understanding the Promise and Challenges of Diverse Membership and Deep Knowledge Integration
Another complexity facing the science of team science is to understand and address the communication and coordination challenges emerging from the first two features that pose challenges for team science introduced in Chapter 1—high diversity of team or group membership and deep knowledge integration. The challenges are especially great in transdisciplinary projects that may have multiple scientific and societal goals and require high levels of knowledge integration across disciplines and professions (Frodeman et al., 2010). Thus, a critical issue for the science of team science involves examination of the integrative processes and outcomes in disciplinarily heterogeneous science teams and how they lead to scientific innovations. This understanding is needed whether the project aims for “translational” innovations that are more immediately applicable or more fundamental scientific knowledge.
Establishing Reliable, Valid, and Consensual Criteria for Evaluating Team Science Processes and Outcomes
Evaluating the processes and outcomes of science teams and groups is particularly challenging because of their multiple goals. As the research focus of the science of team science shifts from small, short-term science
teams to larger, more enduring organizational and institutional structures, the goals of a project and the criteria for judging its success vary accordingly. Whereas the primary goals of small teams may entail the creation and dissemination of new scientific knowledge, larger team science centers and institutions often encompass broader goals. Reflecting their multiple goals, large organizational structures require broad metrics to evaluate their effectiveness. Such metrics may include assessments of the extent to which the smaller team science projects they administer bring about intellectual innovations in the near term, and the extent to which the organization is able to coordinate and integrate across projects to translate these near-term scientific findings into new technologies, policies, and/or community interventions (i.e., scientific and societal returns—see Chapter 9 for further discussion).
These higher-level organizations and institutions (e.g., a research center or institute) must be responsive to the scientific and translational priorities embraced by their community and governmental funders, whereas these priorities may be much less salient to individual scientists working on individual projects (Winter and Berente, 2012). Thus, an important concern of the science of team science field is to develop evaluative criteria that are appropriately matched to the respective goals and concerns of the teams, organizations, institutions, funders, and community groups that have a stake in the foci, processes, and outcomes of large programs of team science research. Scholars in the science of team science are concerned with the relative efficacy of alternative team science funding mechanisms and the development of criteria for evaluating the returns on investments in team science projects—questions that have not been explicitly addressed in earlier research on non-science teams (Winter and Berente, 2012).
The field is also increasingly concerned with articulating appropriate criteria for measuring the potential (ex-ante) and achieved (ex-post) outcomes of science teams and larger groups, including those that focus within a single discipline and those that cross disciplines (Holbrook, 2013; Jordan, 2013; Stokols, 2013). In particular, a growing number of science teams and groups have transdisciplinary goals, seeking to achieve scientific advances by not only integrating, but also transcending multiple disciplinary perspectives and to apply the resulting scientific advances (Croyle, 2008; Crow, 2010; Klein, 2010). In response to this trend, the field is concerned with identifying reliable, valid, and consensually agreed-upon criteria for judging the success of such transdisciplinary projects relative to those that are uni- or multidisciplinary (Frodeman et al., 2010; Pohl, 2011).
As a first step toward developing such criteria, the field must develop measures of the processes leading to effectiveness. As teams and groups develop and move through their phases of scientific problem solving, their interactions will change, and the field must identify how to measure these
team processes. Such measures will aid understanding of how team processes are related to the multiple goals of transdisciplinary team science projects. Achieving this understanding requires articulation of a comprehensive, multimethod measurement approach that includes, but is not limited to, bibliometric indices, co-authorship network analyses, experts’ subjective appraisals of team science processes and products, and surveys and interviews of team science participants. Particularly challenging is the measurement of deep interdisciplinary knowledge integration (Wagner et al., 2011), but there are new methods and measures that appear promising as discussed in Chapter 9. Such efforts to measure team processes are often more daunting than developing evaluative criteria to measure team outcomes in other settings. As part of this measurement challenge, the field needs to more clearly differentiate the processes and outcomes of unidisciplinary, multidisciplinary, interdisciplinary, and transdisciplinary science teams.
An essential first step in the process of establishing evaluative criteria is to gain access to practicing scientists to study their interactions and innovations. Although some funding agencies and scientists themselves resist providing such access, it is critical for advancing the science of team science. For example, more than a decade ago, the Institute of Medicine (1999) produced a groundbreaking report on patient safety and errors in health care. As a result, researchers began to gain access to health care settings, illuminating the relationship between medical teams’ processes and patient outcomes and identifying strategies for reducing errors and improving patient safety (e.g., Edmondson, Bohmer, and Pisano, 2001). Providing researchers access to science teams embedded in their research contexts promises similar benefits.
Focusing on Translational and Educational as Well as Scientific Goals
Finally, the science of team science field is concerned with not only research but also translation of the research to improve practice (Spaapen and Dijstebloem, 2005; Stokols et al., 2008a). The translational goals of the field include
- using the research findings on team science to improve community and societal conditions (e.g., through the development of improved clinical practices, disease-prevention strategies, public health policies);
- applying research findings from evaluations of large team science research projects to improve future scientific teamwork and designing organizational, institutional, educational, and science policies
-
and practices to promote effective team science (see further discussion in Chapters 3–9); and
- developing education and training programs and resources to enhance students and scholars’ capacity for effective scientific collaboration in their future or current team science endeavors (Stokols, 2006; COALESCE, 2010; Klein, 2010; National Institutes of Health, 2010; National Cancer Institute, 2011; Vogel et al., 2012; see also Chapter 5).
A Complex Adaptive Systems Approach
Researchers have begun applying the methods and perspectives of complexity science to help understand and address the communication and coordination challenges of team science.
Complexity science uses computer simulations to study “complex adaptive systems,” which are systems made up of multiple parts that continually interact and adapt their behavior in response to the behavior of the other parts (Holland, 1992). By modeling such systems, researchers seek to understand how the aggregate behavior of the system emerges from the interactions of the parts, integrating multiple levels of analysis to build a more thorough understanding of phenomena. For example, Liljenström and Svedin (2005) described a complex adaptive system as a network of non-linear interactions within an open system, which produce a form of self-organization and emergence. It may be relevant to draw on complexity theory to bound the levels of analysis and address the theoretical and measurement issues present in team science environments. Organizational scientists refer to team effectiveness as “emergent,” because it originates in the thinking and behaviors of individual team members and is amplified by team members’ interactions (Kozlowski and Klein, 2000). Kozlowski et al. (2013) have studied emergent collaboration, and Kozlowski et al. (in press) have examined knowledge emergence in decision-making teams, relevant to the challenge of deep knowledge integration in interdisciplinary and transdisciplinary science teams (Kozlowski and Klein, 2000; see further discussion in Chapter 3).
By virtue of their multiple levels of scale (individual, team, organizational, multi-institutional) and many different actors with various motivations and priorities, science teams and groups can display the major characteristics of a complex adaptive system as described by Hammond (2009). Börner et al. (2010) called for a multilevel systems perspective to advance the science of team science. This approach would include macro-level analyses to help understand broad patterns of collaboration within and across scientific fields (e.g., Klein, 1996), meso-level analyses to understand the social and group processes arising during collaboration in
science teams and groups (e.g., Fiore, 2008), and micro-level analyses to understand the individuals that comprise the science teams (e.g., their education and training, their motivation). Similarly, Falk-Krzesinski et al. (2011) cautioned that “sequential process models could not adequately capture the complexity inherent in SciTS [the science of team science] and may even be misleading” (p. 154). They argued that a systems view is more appropriate as it can help better account for interdependence and the iterative relationships among the components of science teams and the contexts in which they operate.
OTHER CONTRIBUTING FIELDS OF RESEARCH
Many other fields of research in addition to the science of team science and the research on groups and teams contribute to an understanding of team science and how to increase its effectiveness. These include social studies of science (e.g., Galison, 1996), science and technology studies (Pelz and Andrews, 1976), history and philosophy of science, cultural anthropology, and organizational and management studies (e.g., Kellogg, Orlikowski, and Yates, 2006), as well as interdisciplinary studies, information science, the humanities, and program evaluation research. A detailed examination of the contribution of these fields is beyond the scope of this report, but we provide some examples of relevant work in this section.
Sociologists and economists have examined the internal and external forces motivating individual scientists. For example, sociologist Robert Merton (1968) found that well-known scientists were given disproportionate credit for collaboratively authored publications, increasing their visibility while reducing the visibility of less well known contributors. Social scientists continue to study how credit and rewards are allocated when scientists collaborate (e.g., Furman and Gaule, 2013; Gans and Murray, 2015) revealing tensions that affect scientists’ willingness to join science teams and groups (see Chapter 8 for further discussion).
Anthropologists and sociologists have conducted in-depth studies of scientific laboratories in the life sciences, high-energy physics, and other disciplines (e.g., Latour and Woolgar, 1986; Knorr-Cetina, 1999; Owen-Smith, 2001; Hackett, 2005). Cognitive scientists have also conducted studies of scientific work in particular settings, while psychologists have examined the role of scientists’ personality characteristics and other factors in supporting scientific creativity and productivity (e.g., Simonton, 2004; Feist, 2011, 2013). Building on studies focusing on individual scientists, recent research has begun to explore collaborations between scientific institutions (e.g., Shrum, Genuth, and Chompalov, 2007; Garrett-Jones, Turpin, and Diment, 2010; Bozeman, Fay and Slade, 2012; see Chapter 8 for further discussion).
In this chapter, we have described several fields that contribute to understanding how to improve the effectiveness of team science. This report draws heavily on the robust literature from research on groups and teams and on the body of research emerging from the science of team science. We have described the interdisciplinary and multilevel orientation of the science of team science and outlined several of its distinctive challenges and concerns. Many other fields contribute to the committee’s understanding of the effectiveness of team science, but are beyond the scope of this report, including social studies of science, organizational and management studies, industrial-organizational and cognitive psychology, science and technology studies, interdisciplinary studies, communications and information science, the humanities, and program evaluation research.
