Source-linked AI summary
What Large-Scale Publication and Citation Data Tell Us About International Research Collaboration in Europe: Changing National Patterns in Global Contexts
Marek Kwiek
TL;DR
This study examines the growth and changing patterns of international research collaboration in Europe from 2009–2018 using Scopus-indexed publications. It finds that European publication growth is driven by internationally co-authored papers, while individual scientists’ collaboration choices are shaped by prestige, recognition, and funding incentives.
Problem
The study addresses how and why international research collaboration has increased in Europe and how this internationalization relates to publication and citation patterns.
Method
The study analyzes international co-authorship and citation distributions among 5.5 million Scopus-indexed articles published from 2009–2018, including 2.2 million involving international collaboration.
Results
2.2 million EU-28 articles involved international collaboration, and internationally co-authored papers accounted for European publication growth while their field-normalized citation impact was above the global average in almost all European systems.
Takeaways & Limitations
Individual scientists’ willingness to collaborate internationally is central to advancing IRC, with choices motivated by academic prestige, scientific recognition, and access to research funding.
Takeaways & Limitations
IRC entails costs including coordination problems, information overload, unclear responsibilities, and communication issues, while academic peripheries may struggle to maintain research infrastructure.
Abstract
from arXiv · showhide
This study analyzes the unprecedented growth of international research collaboration (IRC) in Europe during the period 2009-2018 in terms of coauthorship and citation distribution of globally indexed publications. The results reveal the dynamics of this change, as growing IRC moves European systems away from institutional collaboration, with stable and strong national collaboration. Domestic output has remained flat. The growth in publications in major European systems is almost entirely attributable to internationally coauthored papers. A comparison of trends within the four complementary collaboration modes clearly reveals that the growth of European science is driven solely by internationally co-authored papers. With the emergence of global network science, which diminishes the role of national policies in IRC and foregrounds the role of scientists, the individual scientists willingness to collaborate internationally is central to advancing IRC in Europe. Scientists collaborate internationally when it enhances their academic prestige, scientific recognition, and access to research funding, as indicated by the credibility cycle, prestige maximization, and global science models. The study encompassed 5.5 million Scopus-indexed articles, including 2.2 million involving international collaboration.
1. Introduction
The paper examines Europe’s exceptional growth in international research collaboration, defining IRC through international co-authorship and asking how it relates to research output, networks, and citation impact. It emphasizes both policy support and scientists’ pursuit of reputation, resources, and strong international partners.
- Argument: The paper argues that collaboration with the best peers, regardless of location, is the primary driver of IRC growth in Europe.This claim complements the paper’s account of policy support by foregrounding scientists’ partner-selection choices.
- Concept and scope: IRC is operationalized as publications co-authored by scientists affiliated with institutions in at least two countries.The study treats research internationalization as an outcome measured through co-authorship rather than as a process.
- Research questions: The study analyzes Europe’s 2009–2018 growth in international co-authorship and citation distribution using globally Scopus-indexed publications.Its research questions address IRC’s contribution to output growth, major country-level networks, and field-specific citation premiums.
- Policy context: European policy has strongly promoted and funded IRC through national and EU programs, with many funded projects involving international collaboration.The ERC budget for 2014–2021 was 13.1 billion euro, while 41.7 billion euro of the 2007–2013 Framework Program budget supported about 26,000 projects.
- Policy context: IRC has become a metric of excellence and quality in the European Research Area and an increasingly important criterion for academic promotion.Major European excellence initiatives have promoted IRC as a key goal while offering concentrated funding.
- Global partnerships: Scientists from leading European universities predominantly seek international co-authors from leading universities globally, producing highly cited joint papers.The paper frames this pattern as “excellence seeks excellence.”
2. Literature review
The literature presents IRC as increasingly shaped by scientists’ incentives and self-organization within global networks, while recognizing substantial coordination and infrastructure costs. Academic recognition, funding access, and prestige-seeking help explain participation, but collaboration creates trade-offs for individuals and national systems.
- IRC and academic careers: Scientists increasingly seek IRC because it can enhance academic recognition and provide better access to research funding.The credibility cycle links prestigious publications to recognition, grant-based funding, and further research outputs.
- Prestige and incentives: Prestige maximization models describe IRC as increasingly important for individual and institutional success, with universities acting as prestige maximizers.The model emphasizes the financial and reputational implications of international collaboration.
- Drivers of collaboration: The literature attributes the extent of IRC more to scientists’ values and predilections than to institutions, disciplines, governments, or agencies.Individual decisions are shaped by available resources, research environments, and trade-offs among collaboration modes.
- Global network science: Global science models portray scientists as increasingly autonomous actors who select collaborators across borders within networked science.The literature describes a shift away from national systems toward scientist-centered networks and bottom-up self-organization.
- Network formation: Preferential attachment helps explain why highly productive and reputable scientists attract additional international collaborators and resources.International networks form around key people who offer knowledge, resources, or both.
- IRC: advantages and costs: IRC advantages must be weighed against national infrastructure risks and individual coordination costs.Overextended collaboration can produce information overload, unclear responsibilities, and communication issues, while academic peripheries may struggle to maintain research infrastructure.
3. Data sources and methodology
The study uses Scopus and SciVal data to examine collaboration patterns across 24 EU member states from 2009 to 2018. It compares international collaboration with institutional, national, and single-authorship outputs.
- Data sources: Data retrieved from Scopus via SciVal between October 20–25, 2019 cover 24 EU member states from 2009 to 2018.Scopus is used as a database of abstracts and citations from peer-reviewed literature and as an overview of world science.
- Methodology: The analysis compares four complementary collaboration types: international, institutional, national, and single authorship.Institutional collaboration has all authors affiliated with one institution; national collaboration has multiple institutions within one country; single authorship is non-collaborative.
4. Results
European research growth from 2009–2018 was driven by expanding international collaboration, while national collaboration remained stable and institutional collaboration declined. Internationally co-authored papers also showed strong citation advantages, especially in EU-13 systems and selected fields.
- IRC, total national output, and system size in Europe: Lower international collaboration is positively correlated with lower higher-education research expenditure, but IRC is only weakly related to national output or researcher numbers.The correlations with publication output and research personnel were negligible at R2 = 0.1 and R2 = 0.06; for the top 100 nations, output correlation was R2 = 0.21.
- Changing collaboration patterns: IRC increased across Europe, while national collaboration remained stable and institutional collaboration and single authorship declined.IRC exceeded 50% in 2018 in all but Croatia, Poland, and Romania; national collaboration decreased by only 0.5 percentage points across the EU-28.
- IRC as the major driver of publication growth in Europe: 46.0% in EU-15 and 30.9% in EU-13: total annual research output rose by these amounts, with major-system growth almost entirely attributable to internationally co-authored papers.Domestic output remained almost flat, while internationally co-authored articles increased steadily; the pattern was confirmed for France, Germany, and the United Kingdom.
- IRC and networks: major partnership countries: The USA was the dominant collaboration partner for most European countries, although geographical, linguistic, and historical ties shaped some pairings.Among the top three European partnership ranks, FWCI was highest for France–Netherlands, Italy–Netherlands, and Belgium–United Kingdom collaborations.
- Field differentiation of international collaboration premiums: Internationally co-authored papers increased FWCI for Poland, Germany, and their top partners, with citation premiums generally higher in EU-13 than EU-15 systems.For all fields combined, the reported increases were 303% of baseline in EU-13 countries versus 159% in EU-15 countries; field-specific premiums reached 1,500% in Bulgarian humanities.
5. Discussion and conclusions
Europe was the global leader in international research collaboration from 2009–2018, with internationally co-authored papers driving publication growth and generally exceeding the global citation average. The findings also highlight a widening EU-15/EU-13 gap and emphasize funding, reward structures, and scientists’ choices as central to future collaboration.
- Europe’s global IRC position: 2.2 million EU-28 articles involved international collaboration during 2009–2018, compared with 1.4 million in the US and 0.7 million in China.In 2018, EU-28 co-authors participated in 57.4% of globally published internationally collaborative articles.
- Europe’s global IRC position: 45.7% of EU-28 articles involved international collaboration, compared with 40.8% in the US.Ten countries had at least one international author on six out of ten articles.
- Citation visibility: The field-normalized citation impact of internationally co-authored papers was above the global average in almost all European systems.FWCI compares received citations with the expected world average for the relevant field, publication type, and year.
- Uneven European patterns: The EU-15/EU-13 research-internationalization gap reflects long-term CEE isolation from global science networks, severe underfunding, and limited institutional pressure to publish internationally.The paper identifies a basic public-research-funding threshold as necessary for IRC that CEE countries have not reached over the last three decades.
- Drivers and implications: In global network science, individual scientists’ willingness to collaborate internationally is presented as key to advancing IRC, especially in CEE countries.The paper links scientists’ choices to academic prestige, scientific recognition, research funding, and reward structures prioritizing internationalization.
- Drivers and implications: Fine-grained field- and institution-level analysis is especially relevant for EU-13 countries, which stand to benefit most from international collaboration and enhanced visibility.The study recommends using detailed European publishing trends to identify fields more or less positively affected by IRC.