Global science for global challenges: the landscape of international scientific collaboration

Global science for global challenges: the landscape of international scientific collaboration

Introduction

The challenges we face, from global pandemics and climate, to security and justice, are increasingly global, complex and dynamic. While the COVID-19 pandemic showed us what can be achieved when scientists are able to collaborate, the economic, social and political barriers to effective international scientific collaboration on many global challenges remain significant, and in many cases are growing even greater. Without deliberate strategies to overcome barriers to collaboration, we risk a knowledge and innovation deficit where the evidence needed to inform strategic decision-making, develop the technologies to address urgent problems, and generate innovations to reduce inequity are unavailable.

What would those strategies look like? The answer is not obvious, nor will there be only one. International scientific collaboration is multi-faceted, highly interconnected and constantly evolving. It would clearly be helpful to identify the factors and dimensions that help or hinder international scientific collaboration. Yet, scientific collaboration is hard to measure, especially in ways that provide meaningful feedback to decision-makers and leaders on the effectiveness of policies or processes designed to support it. While any effort to capture the scope and effectiveness of international scientific collaboration for generating knowledge and innovation risks becoming a fool’s errand, the stakes are too high not to try.

We present a snapshot of the global landscape of scientific collaboration. Our work, informed by a series of multi-stakeholder dialogues combined with original quantitative analysis using publicly accessible data from over 130 countries, explores the dimensions, tensions and knowledge gaps around international scientific collaboration. A companion piece (here) outlines the steps needed to support and secure international scientific collaboration moving forward.

What is needed for successful collaboration?

Individual scientists are often highly motivated to collaborate internationally. Collaboration allows access to study sites, tools, or expertise that are unavailable domestically. Additionally, collaborations often result in increased impact and prestige, as papers with more international co-authorships tend to receive higher citation rates compared to those produced by scientists from the same country.

Motivation is one thing; action is another. International scientific collaboration is multi-faceted, highly complex and dynamic. Understanding the landscape of factors that help or hinder it is a necessary first step to developing durable, resilient and genuinely transformational collaborations. Our stakeholder discussions identified five core pillars underpinning successful international scientific collaboration (Figure 1).

The extent to which any one pillar impacts collaboration depends on a spectrum of choices made by individuals, governments, and businesses, as well as the broader technological, cultural, regional, and political context in which collaboration happens. Nevertheless, it was possible to establish broad consensus that successful, resilient collaborations require alignment among these five pillars.

Which countries are collaborating?

Capturing a global snapshot of collaboration and the forces that shape it is not a straightforward task, as there is no single widely accepted metric to describe collaboration. In part, this is due to the diverse ways in which collaborations happen: across geographic regions, over different time scales (short-, medium- and long-term projects), at different levels of formality, and broad ranging funding models (public and private funders, awarding bodies that are institutional, national or international), not to mention different disciplines, institutions, and sectors.

To consider international science collaboration, we collated data available on international co-authorships on academic publications, which is arguably the most widely accepted metric in the literature (Khor & Yu 2016[1] ; Nature Index).

The first thing to note from the analysis is that some countries collaborate far more than others. Predictably, variation in collaborative output tracks both the overall population size and economic size of the country, at least when measured as gross domestic product (GDP). The US, for example, produced over 1.4 million internationally co-authored publications in 2020 (data downloaded from SciVal) and China over 800,000, whereas low- and middle-income countries produce orders of magnitude fewer. For example, South Africa generated 75,691 papers while Namibia produced 2,029.

The fraction of publications that are from international co-authorships is interesting, since this gives an impression of the extent to which a country’s scientific output relies on international collaboration. Figure 2 shows how the proportion of publications involving international co-authorships changes as a function of total domestic investment in research and development data (measured as government expenditures on R&D, or GERD for short).

Two things are worth noting. One is the downward curving line which suggests that, on average, countries investing least in domestic R&D tend to collaborate internationally relatively more than countries with higher R&D investments. Low investment in domestic R&D increases the likelihood domestic science relies on intellectual and monetary subsidies from other countries. If this situation comes at the cost of de-emphasizing domestic research priorities, there is a real risk that evidence gaps become reinforced, rather than repaired, by international collaboration.

The other is the level of variability of the data. Countries vary in the extent to which they collaborate internationally across all levels of the R&D investment spectrum. Clearly, there can be many other factors influencing either collaboration itself, the total output of the country, or both. Identifying precisely which factor or combination of factors play the most influential roles in advancing or hindering international collaboration in specific contexts is a pressing issue.

The importance of trust

In all our discussions, the one theme every participant emphasized was trust. Trust forms the bedrock on which effective collaborations are built and without it, collaborations are unlikely to be initiated or sustained in the long run.

The importance of trust for international scientific collaboration is highlighted by the World Values Survey (WVS), a global network of social scientists who have been surveying people’s attitudes and values in nearly 100 countries since 1981. The WVS includes a question on the extent to which people of other nationalities can be trusted. Figure 3 shows how the fraction of international co-authorships on publications, our measure of international collaboration, changes as a function of the answers to this survey question.

What is most notable is how the fraction of co-authorships increases as levels of trust increase in high income countries (gold dots and line). This likely reflects strong trade and diplomatic relationships among high income countries, together with their more substantive investments in R&D. Again, the variability overall is high, and especially so for low- and middle-income countries where there is no apparent relationship between collaboration and trust, perhaps reflecting a heterogeneous range of factors such as regional disparities, geography, and more modest investments in research capacity.

There are many gaps in the data around international scientific collaboration

Our analyses of how GERD and trust relate to international co-authorships reveals some intriguing trends but has limitations. We have examined just two correlates of one metric, after all, and we have clearly missed some other important factors that play a role in facilitating collaboration internationally. These factors include those related to mobility, such as visa processing times for visiting researchers, how open a country is to others, and the extent to which countries commit to open science and data sharing, among others. The scatter around the regression lines suggests these other factors could be important. While we did our best to collate as many additional factors as we could, on the advice of our stakeholders, the data were only reported sporadically and often in an uncoordinated way. In short, there were too many gaps to provide a clear picture of the landscape of international scientific collaboration.

Another limitation is our use of international co-authorships to measure collaboration across borders. This metric is just one dimension of collaboration focused by design on academic research. It does not, for example, count collaborations through training programs or projects in the private sector that do not result in academic publications. While it is possible in principle to gather other metrics to reflect these dimensions of collaboration, for example by tracking international students hosted by a country or the nationality of joint patent-holders, again there are often more gaps than records at the national level, at least in the public sphere. In our experience, no single metric captures the complexity of how collaboration happens globally. If we are serious about getting a handle on how well countries collaborate with each other, we will need to commit to producing fuller datasets for a wider set of metrics. 

Summary

Our multi-stakeholder discussions revealed tensions around international scientific collaboration ranging from the familiar, like a lack of funding and prolonged visa processing times, to the more pernicious, such as perceived national security threats and the broader geopolitical retreat from multilateralism. Our quantitative analysis underscores the complexity of this landscape; trends that define features of international scientific collaboration among countries with high levels of domestic investment in research and development do not carry over to the rest of the world. Put another way, there is much we still cannot explain and do not understand about the global landscape of scientific collaboration. However, the presence of unknowns must not be used as an excuse for inaction.

Efforts to reduce and mitigate the impact of barriers to international scientific collaboration will need to face this complexity head on rather than ignoring it. The bulk of what counts, and we mean count literally in the sense of what we can measure, as international scientific collaboration happens as a by-product of domestic investment in research. If we are serious about meeting global challenges we will need to make a renewed commitment to scientific collaboration with purposeful, focused attention and investment by both public and private stakeholders. What this commitment looks like, and the barriers to achieving it, is the subject of our second piece (LINK HERE).

Acknowledgements

We acknowledge productive and valuable conversations as part of the World Economic Forum Global Future Council on Scientific Collaboration with the council co-chairs and members, and council managers Greta Keenan, Alice Hazelton and Sam Leakey.

We thank Catalina Claus who undertook data collection and analysis and K A (Michael) Khor for providing access to the SciVal database.