This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American
An almost universal trend in science today is the growth and prominence of large teams and the receding presence of small teams and solitary researchers. To be sure some highly complex scientific research requires massive numbers of participants; the Nobel Prize–winning LIGO experiment that detected gravitational waves, for example, included over 1,000 researchers, as did the group at CERN that confirmed the existence of the Higgs boson (another Nobel-worthy discovery).
Although the motivation for larger teams is understandable, research in social psychology has extensively documented the benefits of small teams, which outperform larger groups on criteria ranging from efficiency to creativity and innovation. To understand the implications of the shift toward large, complex collaborations, therefore, is of fundamental importance for science and scientists; these are increasingly viewed as core engines of breakthrough ideas and innovations.
Large teams have emerged for many reasons. The problems many scientific groups are addressing are complex, requiring integration of diverse expertise. The internet makes it easier to communicate and coordinate among members of larger and dispersed teams. Furthermore, in science, products of larger teams tend to garner more citations than those of smaller teams. And there is evidence showing that NSF is more likely to fund larger than smaller teams, even though they were equally qualified, suggesting additional impetus for larger teams.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
However, recent research shows that the dominance of larger over smaller groups comes at a cost to innovation and creativity. Wang and colleagues analyzed over 65 million papers, patents and software products produced during 1954–2014, to study the implications of team size on productivity. They used an established measure of disruption to assess how much a given work destabilizes its field by creating new directions of inquiry versus builds on existing ideas and designs to address existing problems. They found that relative to larger teams, smaller teams’ work was more “disruptive.”
In other words, large teams solve problems; small ones generate new problems to solve. They found that as size increased from one to 50 members, level of disruptiveness plummeted. Moreover, they found that size itself, rather than confounding factors, such as differences in topic or type of research design, explained the differences. To be sure, large collaborations contribute materially to science and other creative fields by developing established ideas. But, in a world where large teams flourish as small teams diminish, these results suggest that, at the extreme as science is taken over by large teams, ultimately large teams may run out of problems to solve.
These “Big Data” results fit in nicely with decades of social psychology research, which has documented the diminishing returns of team size on all types of productivity, from Ringelmann’s study of adding people on each side of a tug of war to multiple studies of brainstorming. Whether the task is physical or intellectual, the output per capita decreases as size increases. A team needs the requisite capabilities to address the complexity of its task; however, as tasks become more complex and teams become larger, process losses—problems of coordination of information and motivation accumulate exponentially and interact to reduce productivity.
Two major information-coordination challenges that larger teams have to overcome are the common knowledge effect and the phenomenon of getting the floor. As teams increase in size, members’ redundancy increases. Despite the fact that large teams should have greater access to nonredundant information than small ones, the common knowledge effect occurs—members talk about what everyone already knows instead of the unique information known by just one person.
And yet, that unique information could potentially be combined with the unique information known by one other person to reach creative solutions. Getting the floor also becomes more difficult as size increases. The pattern of direct communication in larger teams looks like a Poisson distribution: one, sometimes two members, do the lion’s share of the talking. For example, early attempts to address the information-coordination problems in groups include Simmel’s study of the dynamics of triads and Morino’s sociograms of relationships within a refugee camp outside Vienna in 1916, which formed the basis for the field of network analysis in contemporary sociology.
The motivational challenges as team size increases are myriad: relational loss, loss of social cohesion, and social loafing, among others. Relational loss is the perception of team members that they are working with little support from other team members. Members of larger teams also feel less cohesive, united, connected and interrelated than members of smaller teams. They tend to be less satisfied with the group and less likely to cooperate with each other.
At the same time, they are more likely to conform to group norms, such as social loafing, the tendency of individual group members to contribute less than they would working in a smaller group or alone. It can escalate in large teams as team members view others as free riding and reduce their contributions so as not to be viewed as a sucker. Free riding occurs as teams increase in size because of diffusion of responsibility, which can be a particular problem when goals and responsibilities are not clear.
As a whole, larger teams can lack the developmental maturity of their smaller counterparts. Throughout the lifetime of the group, larger teams tend to look to leaders for direction and motivation, whereas smaller teams frequently progress to periods of intense productivity fueled by trust-based relationships, structure and consensus. It is no wonder that larger teams are building on existing ideas to solve complex problems, but smaller teams are identifying disruptive new ideas that need solving.
Since it seems neither likely nor necessary for the march toward larger teams to be reversed, the challenge becomes how to make large teams or parts of them into small teams. New research in multiteam systems (MTS) may offer useful lessons. A multiteam system is a large team that is a system of smaller teams organized with a structural network, which is usually, but not necessarily, hierarchical. Groups within a multiteam system share at least one common goal and are interdependent with respect to inputs, process, or outcomes.
Effectiveness in these MTS structures requires team members to balance the demands of their own component team, while also allocating resources to the interdependent needs of the higher-order multiteam system. As research begins to identify the best practices with respect to leadership and team identity that facilitate team performance, it will help us understand how to structure teams—both large and small—that are best positioned for innovation.
In sum, science needs disruptive innovations and ideas, and theoretical and practical solutions. Large teams remain an indispensable engine for problem solving. But we must also recognize the key role of small teams, which have been shown not only to disrupt more than their larger cousins, but also to be associated with greater productivity, better communications and a deeper sense of relational unity, all of which could contribute to strong individual and collective performance.