Small research groups play a crucial role in moving forward the frontiers of our knowledge. This is the view of James Evans, professor at the Department of Sociology at Chicago University.
During the 20th century and up until today, research teams have grown in size, as measured by the number of authors per publication. This development is not just true within certain disciplines.
Within sociology, for example, 11 per cent of articles were co-authored in 1935. By 2005, this proportion had increased to more than 50 per cent. Within particle physics, it is not uncommon for articles to have more than one hundred authors.
Investigated the size of teams
Against this background, James Evans and his colleagues have investigated how the size of the research team impacts on the research produced by the group. Their study is based on the analysis of 65 million research articles, patents and software products from the period 1954–2014, and was recently published in the periodical Nature.
The results show that larger research team develop and continue building on existing ideas to a greater extent. Smaller teams produce a greater amount of research that revolutionises science and technology through new ideas.
The work of larger teams continues building on newer findings, while smaller teams investigate promising ideas from older and less investigated research. Group size is here defined as the number of authors of the article, patent or software.
“It surprised me that the result is so universal. That it applies over time and across different research fields,” says James Evans.
How research is cited
In the study, the researchers investigated how disruptive the research is, by analysing how published research is cited. If the citations of a certain article do not refer to the predecessors of the article, then the research in the article is considered as disruptive.
When the opposite applies – if the articles citing a certain article also refer to a large number of its references – then the article is instead considered as developing research.
The study analysed research results from nine research fields (physics, biology, medicine, environmental science, chemistry, agriculture, social sciences, technology and computer science). The link to team size was seen in all areas apart from the last two, for which the main publication source was not included in the analysed material.
The critical point varies
For each field, the researchers identified the point where groups changed from mainly producing disruptive research to mainly producing developing research.
“This point differed between the fields. All research fields are not the same, and we are now further investigating these differences,” says James Evans.
The researchers established that both small and large teams are needed in order for the research ecosystem to flourish. They hope that the results will contribute to changing the current funding system, which they consider rewards risk-minimisation and therefore not potentially disruptive research.
“If research funding bodies want to realise the potential of science, they must increase the funding of the risk-filled spectrum of research, and this will include small teams,” says James Evans.
Greater cohesion in smaller teams
Sven Hemlin, Professor Emeritus of Psychology at the University of Gothenburg, thinks that the study in Nature is well designed, using a large database. It also uses a measurement that is interesting, disruptive research, but which can of course be discussed.
According to Hemlin, it is known that the size of the research team affects the cohesion of the team – it is greater in smaller teams – and that there are more people to talk to in larger teams. When it comes to creativity in research teams, it is known that leadership, team climate and composition are of quite some importance.
“But team size has no importance for creative results, according to previous studies. This study contradicts this in part, as there is an overlap between creative research and what they call disruptive research. Their results indicates that team size has an importance that was not previously known,” says Sven Hemlin.
He also considers that the study in Nature has implications for research policy.
“Both representatives of research policy and researchers have long been of the view that there must be a minimum number of researchers – a critical mass – in order to result in good science. They have quite simply invested in larger teams. But there no support for this producing more innovative and valuable results, it is a myth.”
Other factors too
Merle Jacob, Professor of Research Policy at Lund University, considers that the study in Nature highlights the fact that not all research can be directed via one support format. She is not surprised by the results, but at the same time thinks that they should be taken with a pinch of salt.
“Team size is not the most important factor determining what emerges from the team. Instead, it is about factors such as resource allocation and the importance of having a successful researcher in the team. Team size has become a big issue, as it is something research funding bodies feel that they can influence,” she says.
On the other side of the Atlantic, James Evans establishes that of course there may be other factors, correlated to team size, that explain why small teams are better than large teams at disruptive research.
“We have our suspicions – it looks like there may be a component that has to do with money and risk, and another one to do with hierarchies, for example. We are now investigating this further experimentally,” he says.