Go back

Close colleagues spark bright ideas

                 

Researchers might need to be near their collaborators to break new ground, says John Whitfield

Researchers have long been used to remote working. Projects involving far-flung colleagues and the use of technology—email, video calls, messaging apps and document drops—are commonplace for those in nearly all disciplines.

It should be no surprise, then, that the debate around the post-pandemic return to the office that has raged in many industries has largely passed research by. With solid evidence that international collaborations are on average more highly cited, some have even argued that collaboration has little need of contact.

But many researchers say real-world spaces matter deeply to how research works. Otherwise, those designing buildings and facilities such as London’s vast Francis Crick Institute, a biomedical lab that opened in 2016, wouldn’t put vast effort into creating spaces and routines that encourage mingling, conversation and serendipitous encounters based on such a belief. 

So, do the opportunities opened up by having a whole world of colleagues to choose from really outweigh those of having collaborators close at hand? Researchers are now trying to answer this question—and are coming up with some striking results.

Close friends

In a recent preprint, Yiling Lin of the University of Pittsburgh, together with colleagues, looked at 20 million articles published between 1960 and 2020 by 22.5 million authors in 3,562 locations, lobbing in four million patents from the same period for good measure. They found that the average distance between team members changed over this period from 100 kilometres (Lille to Brussels) to 800km (Stuttgart to Rome).

Lin’s team scored each paper as disruptive or developing, based on the extent to which later papers that cite it also reference the same papers that it cites. Disruptive work, they reason, replaces its predecessors, while developmental studies join theirs. 

As collaboration distance rises, they found, the probability of disruptive work falls. The effect is particularly steep at short distances, suggesting that physical presence is key. Just being in the same time zone isn’t enough.

Using a smaller dataset of papers disclosing author contributions, they suggest that this is because remote work is well suited to technical tasks such as collecting and analysing data, but less good for coming up with new ideas and designing research programmes, which come out of people bouncing ideas off one another. The rise of remote work, they hazard, may be contributing to why, in disciplines such as drug discovery, breakthroughs are getting rarer and more costly in terms of money and effort. 

Post-pandemic pressures

The pandemic, of course, imposed remoteness on people who had previously shared a workplace, and many of them seem to have liked it that way. In another recent study, Daniel Carmody and his colleagues looked at how lockdown changed researchers’ social networks.

Carmody works at the Massachusetts Institute of Technology, and his team’s study reconstructs social ties by mapping the university’s email traffic. When lockdown hit, they found that the most significant change was a drop in the number of weak ties. These are the people at the edge of your networks; because your circles don’t overlap much, they are especially useful as a source of new ideas and opportunities, and without them networks are more likely to ossify and become echo chambers. The advent of hybrid working when restrictions eased restored some of these ties, but they did not return to their former level.

Carmody and his colleagues call for efforts to work out the minimum amount of time in the workplace needed to sustain the benefits of water-cooler chat, and they want employers to work out how to encourage serendipitous encounters between staff.

Lin’s team argues that managers should assign disruptive projects to on-site teams and developmental ones to remote colleagues, and that policymakers should not focus on digital infrastructure at the expense of transport and housing policies that help people come together. 

One could quibble that patterns in reference lists are a broad-brush measure of the impact of research, and there seem likely to be other reasons for falling scientific productivity besides a growing distance between collaborators.

For what it’s worth, both Lin and Carmody produced their work with colleagues in the same institution and with others in Europe. But together, these two studies hint that being in the same place gives a boost both to deep collaborations and to the chance encounters that spread information and spark ideas. 

John Whitfield is comment editor for Research Professional News

This article also appeared in Research Europe