Network structure of scientific collaborations between China and the EU member states
Network structure of scientific collaborations between China and the EU member states
Lili Wang 0 1 2 3 4
Xianwen Wang 0 1 2 3 4
Niels J. Philipsen 0 1 2 3 4
Science 0 1 2 3 4
Network 0 1 2 3 4
Structure 0 1 2 3 4
0 WISE Lab, Dalian University of Technology , Dalian 116024 , People's Republic of China
1 UNU-MERIT, Maastricht University , Boschstraat 24, 6211 AX Maastricht , The Netherlands
2 & Niels J. Philipsen
3 RILE, Erasmus University Rotterdam , 3000 DR Rotterdam , The Netherlands
4 METRO, Maastricht University , 6200 MD Maastricht , The Netherlands
Collaborations between China and the European Union (EU) member states involve not only connections between China and individual countries, but also interactions between the different EU member states, the latter of which is due also to the influence exerted by the EU's integration strategy. The complex linkages between China and the EU28, as well as among the 28 EU member states, are of great importance for studying knowledge flows. Using co-authorship analysis, this study explores the changes of the network structure between 2000 and 2014. Our results show that EU member states with middle- or low- scientific capacities, in particular those who joined the EU after 2000, have been actively reshaping the network of scientific collaborations with China. The linkages between middle- and low- scientific capacity countries have been tremendously strengthened in the later years. The network positional advantage (measured by the degree of betweenness centrality) has shifted from a few dominant nations to a wider range of countries. We also find that countries like Belgium, Sweden and Denmark are in important positions connecting the relatively low-capacity 'new' EU member states with China. The 'new' EU member states-that have relatively low scientific capacity-intend to cooperate with China jointly with 'old' EU member(s).
International collaboration; Integration; EU member states; Centrality
Committed to promoting science and technology and increasing China’s global influence, the
Chinese government has placed great emphasis on international cooperation. In recent
decades, China has gradually established science and technology collaborations with more
than 150 countries and signed cooperation agreements with nearly 90 countries
. In evaluating collaboration results, however, the existing literature has been
mainly focusing on collaborations between China and the U.S
(Wang et al. 2012, 2013; Tang
2013; Tang and Shapira 2011; Suttmeier 2014; Hand 2010)
. Little is known about
collaborations between China and its second biggest partner, the European Union. In recent years,
along with the fast and comprehensive growth of China-EU relations, the collaboration in
S&T between China and European countries has also grown rapidly, which could be reflected
by increased mobility of people (including scientists), collaboration in research projects, and
coauthorship in publications (Li and Chang 2014). Neverthess, the limited research on
scientific collaborations between China and the EU so far has been mainly conducted at the
national level, for instance in relation to the UK
(Bound et al. 2013; Zhou et al. 2013)
In fact, the European countries have a very high national share of international
(Zhou and Gla¨nzel 2010; Hand 2010)
. The EU level of international co-authorship is
about twice that of the United States (Hand 2010). Nevertheless, due to the funding
programmes from the European Commission and the spatial proximity in Europe, the pattern of
intra-EU collaboration is different from that of extra-EU collaboration
(Mattsson et al. 2008;
. In the process of collaborating with China, the interaction between the
heterogeneous 28 member states of the European Union (EU28) is of great interest.
The contribution of this study is twofold. First, from the extra-EU perspective, this
study aims to provide an in-depth understanding on the evolution of collaboration networks
between China and the 28 EU member states. Mapping the change of the network structure
is crucial in advancing future knowledge transfers between China and the EU28. Second,
from the intra-EU perspective, this study offers insights on the interactions between
European countries. By analyzing the network structure of scientific collaborations, we
investigate whether European countries have been (further) integrating while collaborating
with China. If so, to what extent (and in what scope) such integration is happening? Special
attention will be paid to the change of collaboration patterns of ‘new’ member states,1 as
well as the interactions between the ‘new’ and ‘old’ member states.
The remainder of this paper is organized as follows. In ‘‘Backgrounds’’ section
summarizes the background of China-EU cooperation and reviews related literature. In ‘‘Data
and Methodology’’ section provides data and methodology. Analyses and results are
documented in ‘‘Results’’ section and ‘‘Conclusions’’ section concludes.
Cooperation efforts from China and the EU
Along with its rapid economic growth and social change, China has been seeking to
integrate into international structures
(European Commission 1995)
. Aiming at more
1 We use the terminology ‘old’ and ‘new’ EU Member States to refer to, respectively, the countries that
formed the EU15 and those that joined the EU in 2004 or later. For further details, see ‘‘Dynamics of
scientific collaborations’’ section.
bilateral cooperation, China and the EU signed the Science and Technology Agreement in
December 1998, which was renewed in 2004, 2009 and 2014.2 Jointly steered by DG
Research and Innovation and the Chinese Ministry of Science and Technology (MoST), the
European Atomic Energy Community (EURATOM) and China signed an agreement for
R&D Cooperation in the Peaceful Uses of Nuclear Energy (R&D-PUNE Agreement) in
April 2008.3 In order to complement and strengthen further fruitful scientific cooperation,
MoST and DG Research and Innovation signed the Agreement on Implementing the
Science and Technology Partnership Scheme (CESTYS) in May 2009.4
As the largest research funding agency for basic research and application-oriented
research in China, NSFC (National Natural Science Foundation of China) signed the
Agreement of Scientific and Technological Cooperation between Chinese government and
the European Community in 1998, which was renewed in 2009, aiming to launch projects
in specific research areas of common interest. Many other bilateral agreements between
NSFC and the EU countries or various funding agencies in the EU have been reached, e.g.
agreements with DFG (German Research Foundation), ANR (French National Agency for
Research), FWO (Research Foundation—Flanders), STINT (Swedish Foundation for
International Cooperation in Research and Higher Education), FWF (Austrian Science
Fund), NWO (Netherlands Organization for Scientific Research), etc.
The European Union has also made efforts in promoting the cooperation with China.
The EU’s Framework Programmes, including the FP6 (6th Framework Programme for
Research and Technological Development, 2002–2006), FP7 (2007–2013) and Horizon
2020 (2014–2020) programmes are all fully open to international cooperation, and China is
the most important cooperation country in these programmes
Being aware of the fact that China was in the midst of sustained and dramatic economic
change and that China has been seeking integration into international structures, the
European Commission (1995) issued an official document—A long term policy for China–
Europe relationship—listing China’s importance for Europe and stressing that ‘‘Europe
must develop a long-term relationship with China that reflects China’s worldwide, as well
as regional, economic and political influence’’.
Owing to the efforts and cooperation agreements from both sides, scientific
collaborations between China and the European Union have been growing dramatically. The
number of papers co-authored by Chinese researchers and researchers affiliated in the
EU28 has increased from 2500 in 2000 to more than 19,000 in 2014. For most EU member
states, China is the second most extra-EU collaborating country only after the USA in
2016.5 Despite the rapid growth in collaborations, however, little is known about the
structure of collaboration networks between China and the EU, nor the dynamics of how
EU member states interact while collaborating with China.
2 See more at http://eeas.europa.eu/archives/delegations/china/eu_china/research_innovation/st_relations/
3 See also at http://ec.europa.eu/research/iscp/index.cfm?pg=china.
4 The official document is available at https://eeas.europa.eu/sites/eeas/files/s_t_partnership_agreement_
5 Based on data from Web of Science.
International scientific collaborations
Collaborations have been widely recognized as a more and more common phenomenon in
science. Through collaboration, partners can share knowledge, skills, techniques and
(Katz and Martin 1997; Beaver and Rosen 1978; Price and Beaver
. International collaborations have been believed to generate even higher social
impacts than domestic collaborations. Based on a set of papers published in 1995 and 1996,
Gla¨nzel (2001) points out that international co-authorship generally results in publications
with higher citation rates than purely domestic papers. Using a database containing nearly
a half million refereed UK publications,
Katz and Hicks (1997)
find that domestic
collaborations (with authors from the same institution or another domestic institution)
increase the average impact by approximately 0.75 citations while international
collaborations (with authors from foreign institution) increase the impact by about 1.6 citations.
With a set of SCI papers published by European institutes, Narin et al. (1991) find that the
citation received by internationally co-authored papers is twice as high as that received by
papers authored by scientists working at a single institution within a single country.
Nomaler et al. (2013)
stress the factor of geographical distance, pointing out that
collaborations between geographically distant researchers tend to receive higher citations.
In China, the 15-year ‘‘Medium- and Long-Term Programme for Science and
Technology Development’’ issued in 2006 demonstrated that one of the primary goals by 2020
is to have the average cited scientific publications of Chinese authors reach the top 5
worldwide.6 As citation improvement relies on earlier reputation and recognition
, this goal may be difficult to be fulfilled only by domestic publication. In this regard,
publishing jointly with international scholars would be beneficial for Chinese scholars to
advance their recognition in particular outside of China. By tracking corresponding authors
of the joint publications between China and the EU28, Wang and Wang (2017) find that the
academic cooperation between China and the EU has been mainly set up by Chinese
researchers, and in the fast-growing collaborative fields the scores of revealed comparative
advantages have all improved in China.
In the process of collaborating with China, the intra-European collaboration is also of
great importance, which has been embodied by the European Union’s strategic goal to
strengthen regional cohesion in the EU and stimulated work towards an European Research
(European Council 2000)
. In line with the aim of integrating and coordinating
research activities at national and Union level, scientific communities of Western and
Eastern Europe are in particular encouraged to be integrated
(European Commission 2000)
According to the ERA survey, the national research systems of the EU member states have
become more aligned to the ERA priorities, and member states have been increasingly
open to international cooperation
(European Commission 2014)
. Based on scientific
publications between 1998 and 2004,
Hoekman et al. (2009)
find that collaborations
between European regions have been impeded by geographical barriers, and the research
activities in the EU is far from being integrated.
shows that some EU
member states have a relatively strong preference for research partners.
Given the great heterogeneity of European countries, country size and capability should
be taken into consideration in studying collaborations with and within Europe. Lacking of
collaborative partners within the national borders, smaller European countries are expected
6 The official guideline of the ‘‘Medium- ‘Medium- and Long-Term Programme for Science and
Technology Development’’ (in Chinese) is available on the website of MSTC (Ministry of Science and
Technology of the People’s Republic of China) http://www.most.gov.cn/kjgh/kjghzcq/.
to collaborate more internationally
. With a development model,
argues that the share of a country’s internationally co-authored articles is dependent
on the phase of a country’s scientific development. Internationally co-authored
publications are expected to grow very fast in a country’s building-up stage, which could explain
the fast growth of the international collaboration of China during the last two decades.
In measuring research collaborations, several types of indicators are often used. First,
bibliometric analysis of co-publications is regarded as ‘‘one promising approach’’
and Persson 1996)
and it ‘‘provides a window on patterns of collaboration’’
. This approach has been widely applied in examining collaborations at country level
(Beaver and Rosen 1978; Katz and Hicks 1997; Gla¨nzel 2001; Coccia and Wang 2016;
Wang and Wang 2017; Wang et al. 2013)
, at regional/city level (Hoekman et al. 2009), and
institution or author level
(Yan and Guns 2014; Wang et al. 2013)
. Following that,
coinventorship is another often used as an indicator in analysing research collaboration
et al. 2015; Morescalchi et al. 2015; Gao et al. 2011; Hoekman et al. 2009)
Gao et al.
state that co-publications and co-inventions are the main types of outcome of
research collaboration. A third group measures research collaboration by the number of
joint research and development (R&D) projects
(Scherngell and Barber 2011; Scherngell
and Lata 2013; Revilla et al. 2003; Autant-bernard et al. 2007; Hazir and Autant-Bernard
. Joint participations in projects reflect the common R&D activities between partners.
It is worth noting that, as pointed out by Melin and Persson (1996), not all research
collaboration necessarily leads to one type of output (e.g. co-authored papers,
co-inventions or R&D projects). Among these indicators, co-authorship of articles has been widely
accepted to measure collaboration within the academic community
(Melin and Persson
1996; Newman 2004)
Structure of collaboration networks
The structure of collaboration networks reveals the pattern of knowledge exchange and the
mechanism of preferential attachment
(Wagner and Leydesdorff 2005)
. In scientific
collaboration networks, important changes have been predicted to be under way
and Ponomariov 2011)
states that the new ‘‘regional networks are
reinforcing the competence and capacity of emerging research economies, and changing the
global balance of research activity’’
(Adams 2012, p. 335)
Strong ties have been believed to be more important for exchanging knowledge than
(Fritsch and Kauffeld-Monz 2010)
. Lacking of collaboration tendency, scientific
super powers may lose their advantageous position in the global networks
Leydesdorff and Wagner (2008)
find that the global collaboration network has reinforced
the formation of a core group consisting of fourteen most cooperative countries during the
period 2000–2005, and this core group is expected to use knowledge from the global
collaboration networks with great efficiency.
Toward tracking knowledge flows,
Breschi and Lissoni (2003)
argue that connection in
the social network (known as social proximity) is of more importance than closeness in
geography (known as spatial proximity), which is confirmed by
Autant-bernard et al.
: ‘‘social distance seems to matter more than geographical distance’’.
Hoekman et al.
find that the ongoing process of European integration is removing territorial borders
(regional, national, language) on the intensity of research collaboration across European
Scherngell and Lata (2013)
also confirm that geographical distance and country
border effects on the collaborations among 255 European regions gradually decrease over
the period 1999–2006. Following this theory, despite the close distance between some
European countries, the ‘‘active participation in a network’’ is the key for knowledge
exchanges. To this end, exploring the China-EU network structure provides a deeper
understanding on knowledge flows not only between China and the EU, but also between
Data and methodology
There are various methods of measuring research collaborations, as discussed in
‘‘International scientific collaborations’’ section. In this study, we adopt the co-authorship
measurement. Data are collected from Science Citation Index Expanded (SCI-E) and
Social Sciences Citation Index (SSCI) of Thomson Reuters (currently known as Clarivate
Analytics). In our analysis, we focus on the international collaborations at national level.
Affiliation address is used to identify the location of authors.
In exploring international scientific collaborations, one needs to ‘‘go beyond absolute
differences in country sizes and estimate propensities or intensities of collaboration’’
(Luukkonen et al. 1993)
. In this regard, Jaccard and Salton indexes are the two commonly
used measures to capture the relative strength of bilateral connections between countries
(Luukkonen et al. 1993; Leydesdorff 2008; Boschma et al. 2014)
. Another type of
indicator, the Probabilistic Affinity Index (PAI), introduced by Zitt et al. (2000), calculates the
ratio of observed and expected number of links. Similar to PAI, Probabilistic Partnership
Index (PPI) also measures the deviation from expected values of collaborative linkages
(Yamashita and Okubo 2006)
. The difference between PAI and PPI is that the former
measures the expected value from the number of links while the latter is calculated from
the current participants
(see more details in Yamashita and Okubo 2006)
. Both PAI and
PPI methods seem to be extremely sensitive to the level of breakdown of collaborating
(Zitt et al. 2000; Zitt and Bassecoulard 2004; Yamashita and Okubo 2006)
. In our
study examining collaborations between China and multiple European countries at the
same time, we adopt the first set of measurements, i.e. Jaccard or Salton index.
et al. (1993)
states that ‘‘the Jaccard measure is preferable to Salton’s measure since the
latter underestimates the collaboration of smaller countries with larger countries’’.
also suggests that ‘‘Jaccard index is the best basis for the normalization
because this measure does not take the distributions along the respective vectors into
account’’. Considering the size difference between China and most European countries, the
Jaccard index is a more suitable choice for our study.
The collaboration strength index using Jaccard’s measure
(Luukkonen et al. 1993)
be expressed as:
Cij ¼ Pi þ Pj
ði 6¼ jÞ
where COij is the number of co-authored papers between country i and country j; Pi is the
number of total publication by country i; Pj is the number of total publication by country j.
We have further compared the Jaccard index and the Salton index7 in our dataset.
Applying both methods to our data set yields similar results. Given that regional difference
is slightly more pronounced in the results produced by the Jaccard index, we choose this
index in our analysis.
7 Salton index is calculated by CSI ¼ COij= Pi Pj.
Following Eq. (1), calculating the bilateral connections between each pair of the 29
countries (i.e. China and 28 European countries) produces a 29 * 29 collaboration matrix.
We repeat this calculation for different years to derive a set of dynamic matrices. To avoid
exceptional cases in the network evolvement, we use a five-year window scheme. We use
the general structure in 2000–2004 to represent the earlier stage, and the one in 2010–2014
to represent the later stage. Hence the Network evolution is captured by a comparison of
network changes between the two periods, i.e. the period of five earlier years (2000–2004)
and the period of five later years (2010–2014).
In order to examine the change of central players in the network, we calculate the
betweenness centrality for both stages. The betweenness centrality of country i is defined as:
Psk;tðiÞ ðs ¼ country1; country2; . . .country29;
k ¼ country1; country2; . . .country29; i 6¼ s 6¼ kÞ
Psk is the number of shortest paths from country s to country k at year t. PskðiÞ is the
number of shortest paths from country s to country k that contain country i at year t.
In order to keep the degree of betweenness centrality comparable over time, we
normalize the above centrality value to a scale within 0 and 1.
CtnðiÞ ¼ maxðCtÞ
where CtnðiÞ is the normalized betweenness centrality of country i at year t; maxðCtÞ is the
maximal value of betweenness centrality of all the 28 EU countries at year t; minðCtÞ is the
minimal value of betweenness centrality of all the 28 EU countries at year t.
Dynamics of scientific collaborations
During the period between 2000 and 2014, there were in total over 123,800 joint
publications between China and the EU 28, among which the top three collaborating European
countries with China were the UK (42,561), Germany (33,352), and France (20,401),
followed by the Netherlands (10,455), Italy (10,257) and Sweden (9860). These six
countries jointly published with China 107,459 articles in total, accounting for nearly 87%
of all China-EU28 joint publications.
It is noteworthy that the European Union has evolved over the years. Formed by six
countries8 in 1958, the EU has gradually expanded to 28 member states by 2014. Thirteen
countries joined the EU in the 2000s, namely Czech Republic, Estonia, Cyprus, Latvia,
Lithuania, Hungary, Malta, Poland, Slovenia and Slovakia in 2004, Bulgaria and Romania
in 2007, and Croatia in 2013.9 Due to the time difference in joining the EU, different
collaboration patterns have been observed between the established 15 member states and
the 13 new members
(Makkonen and Mitze 2016)
Makkonen and Mitze (2016)
8 They are Belgium, Germany, France, Italy, Luxembourg and the Netherlands. 9 For a detailed overview of EU history and integration, see e.g.(Craig and De Bu´rca 2015, pp. 1–23).
we disentangle the established 15 ‘old’ EU member states10 which joined the EU before
2004 from the 13 ‘new’ ones.
Given that country sizes and research capacity vary greatly among the European
countries, propensities toward international collaborations are also different across nations.
In terms of absolute co-publication counts, regions with large research outputs are believed
to collaborate more
(Hoekman et al. 2010)
. However, to capture the relative strength of
links between countries, it is necessary to eliminate the country-size effect
et al. 1993)
. As explained in ‘‘Data and methodology’’ section, the propensity of
collaboration in this study is measured by intensity instead of absolute publication numbers.
Before the network analysis, we first examine the propensity of international
collaborations, with a connection of scientific capacity.
To reflect the general level of international collaboration strength between one country
and the other 28 partners, the mean was taken after calculating the Jaccard collaboration
strength (Eq. 1) in the matrix. Figure 1 displays the averaged collaboration strength (which
is called collaboration intensity in the figure) as well as the scientific capacity of each
(proxied by the total publications during the period of 2000 and 2014)
During the studied 15 years, the UK and Germany published over 1.5 million articles,
which was similar to the amount of publications by Chinese researchers. In the EU28, the
old members, such as the UK, Germany, France, Italy, Spain, the Netherlands and Sweden,
were in the top quartile, representing the scientifically strong countries. Located in the
second and third quartile, countries between Poland and Slovakia can be defined as the
middle-level-capacity countries. Latvia, Luxembourg and Malta had the fewest
publications during the studied 15 years.
Figure 1 depicts the collaboration propensity (i.e. intensity of collaborations between
China and the EU28, measured by the Jaccard index) and research capacity (i.e. number of
total publications at national level). We label the 13 new EU members in light blue
squares. To offer a dynamic view, the collaboration intensity is provided for two time
periods, i.e. the earlier stage between 2000 and 2005 (blue line) and the later stage between
2010 and 2014 (red line).
For China, the intensity of collaborations with the whole EU has increased from 0.04 to
0.08. In European countries, as shown in Fig. 1, the research capacity and collaboration
intensity are considerably heterogeneous. The EU countries with strong research capacity
(proxied by the number of total publications at national level) are the old EU members,
such as UK, Germany, France, Italy, Spain, the Netherlands and Sweden. In these
countries, the intensity of collaborations with China increased almost equally, from on average
0.7 to 5%. However, countries in the 2nd, 3rd and part of the 4th quartile, which represent
middle- and low-level-capacity countries, have increased their collaboration intensity
most. Austria, Greece and Portugal are the three old EU members who have advanced their
collaboration strength remarkably, to above 8% by the later stage. Poland is the only new
EU member state which is located in the 2nd quartile in terms of scientific capacity.12 As
an old EU member, Ireland’s scientific capacity is positioned in the third quartile, and its
collaboration intensity stays relatively low in the second stage. All the new member states,
10 This includes Austria, Belgium, Denmark, Finland, France, Germany, Greece, Italy, Ireland,
Luxembourg, Netherlands, Portugal, Spain, Sweden and UK.
11 Scientific capacity can be evaluated by different indicators, such as total R&D input, R&D per capita or
various types of scientific output. Here we use the volume of academic publications to represent the level of
12 Note again that this is measured only by the volume of total publications.
1st quar le
total publica ons_2000-14
except Latvia and Malta, have improved their collaboration intensity to around 10% or
even more. Being an old EU member, Luxembourg’s Jaccard index remained low in both
stages. This indicates that, together with Latvia and Malta, Luxembourg was isolated in the
collaboration network with China.
The change of centrality of the collaboration network
In the process of collaborating with China, due to the geographical location and historical
factors, etc., European countries tend to have preferential partners within the EU.
Betweenness centrality is an important structural attribute in co-publication networks, and
nodes with a high centrality degree play the roles of being brokers or gatekeepers to
connect the nodes and sub-groups
(Freeman 1978; Abbasi et al. 2012)
. The betweenness
centrality helps to understand the position of a country in the knowledge flow. A higher
centrality represents a more important as well as more advantageous position of this
(Leydesdorff and Wagner 2008)
. The EU countries with a high level of centrality
are regarded as the important nodes in connecting China and other EU member states.
Figure 2 provides the degree changes of the 28 European countries over the years. In
Fig. 2, the y axis represents the degree of betweenness centrality and the x-axis represents
the logarithm value of the number of articles jointly published with China.
In the earlier stage 2000–2004, illustrated in Fig. 2a, the superpower Germany
dominated the central position, with the highest degree of betweenness centrality. Following
that, other high-capacity countries like the UK, France, Italy, Spain and the Netherlands
also had a relatively high level of betweenness centrality. In the later stage 2010–2014,
presented in Fig. 2b, the betweenness centrality of the UK and France had increased
tremendously. There was, however, a dramatic decline for the betweenness centrality of
Germany. In fact, Germany was one of the only two countries with the centrality
decreasing, while the other country was Cyprus. As shown in Fig. 2b, in the later stage,
France, the UK, and Germany had similar levels of betweenness centrality. Figure 2c
captures the changes of countries between these two stages. From the earlier stage to the
later stage, almost all the countries have increased both the volume of collaborated
publications and the betweenness centrality, which can be reflected by the moving trend
from the lower left to the upper right in the scatter plot in Fig. 2c.
Evolution of collaboration networks
In order to understand the pattern of knowledge exchange in the process of China-EU28
scientific collaborations, this section maps the collaboration network and detects its
changes. Special attention will be paid to the change of collaboration patterns of new
member states (after they joined the EU), as well as the interactions between the new and
old member states.
Figures 3 and 4 provide the collaboration network structure in two different stages, an
earlier stage of 2000–2004 and a later stage of 2010–2014. Publications covered in the
earlier stage are related to research activities before the thirteen member states joined the
EU.13 In the network figures, new EU members are labelled in light blue squares.
As shown in Fig. 3, there are two main clusters in the earlier stage. Led by the high
research capacity countries, e.g. Germany, the UK, France, Spain, the Netherlands, etc., a
cluster consists of 14 countries plotted at the lower level of the figure. This cluster consists
of mainly pre-existing EU members. Three countries—Slovenia, Poland and Czech
Republic, who joined the EU in 2004, are also in this cluster. On the top of the figure, there
13 Although 10 countries (Czech Republic, Estonia, Cyprus, Latvia, Lithuania, Hungary, Malta, Poland,
Slovenia and Slovakia) joined the EU in May 2004, the publications indexed in 2004 reflect research
activities that were carried out some time before the publication date, i.e. before they joined the EU.
is another cluster mainly formed by new member states. Three old EU
members—Portugal, Finland and Ireland—also belong to this group. In this cluster, Cyprus, Bulgaria,
Hungary and Romania in particular are strongly connected with each other. Though
belonging to the same cluster, Slovakia, Croatia, Lithuania and Estonia are hardly
connected with others.14 Luxembourg is isolated from either of the main clusters.
At the later stage (Fig. 4), after the 13 new members have also joined the EU, the main
clusters have developed from two to three. The connections between the high capacity old
EU members (i.e. Germany, the UK, France, Italy, Spain, The Netherlands) are sparse.
Interestingly, however, many of the new EU member states have tremendously increased
their collaboration links with each other.
Another interesting observation from Fig. 4 is that some old EU member states (i.e.
Belgium, Sweden, Denmark and the Netherlands) are located in very important positions in
the network. It has been shown in ‘‘The change of centrality of the collaboration network’’
section (Fig. 2) that these countries have increased their betweenness centrality remarkably
in the later stage. If these nodes are removed from the recent network, the majority of the
new EU members can hardly be connected with China any more. In other words, these old
14 In order to show the main linkages, network filter has been applied. Thus the weak edges are not invisible
in the network figures.
EU member states act as the brokers linking up China and the new EU member states.
Hence, knowledge exchange between China and these new EU member states—that are
middle and low capacity countries—crucially depend on the brokers who are connecting
Given that all the publications from our dataset are those jointly published with China,
the figures reveal that the small countries (on the top of the figures) collaborate with China
together with the brokers, e.g. Belgium, Sweden, Denmark and the Netherlands. The
figures also illustrate that there is a very strong clustering effect on the top of Fig. 4, which
consists of countries with middle- or low- research capacity,15 including mostly new EU
member states and a few old member states. On the top left, Finland and Belgium belong to
the cluster which includes the new members such as Bulgaria, Cyprus, Estonia, Lithuania
and Croatia. On the right top, the cooperation cluster includes five old EU members
(Sweden, Denmark, Austria, Greece and Portugal) and six new EU members (Hungary,
Czech Republic, Slovenia, Slovakia, Romania and Poland).
Comparing the two networks (Figs. 3, 4), judging from the link strength between
countries, it shows that the knowledge integration in the new EU member states have
increased remarkably. This signals that joining the EU has exerted an obvious influence on
scientific integrations between these new member states. However, there are still some
15 Research capacity is proxied by the number of publications (see more in ‘‘Dynamics of scientific
collaborations’’ section, Fig. 1).
countries, e.g. Luxembourg and Ireland, which are relatively isolated. To facilitate a strong
integration, special attention needs to be paid to the isolated countries.
Middle-sized countries are the most dynamic players in the process of network
evolution. While the superpowers like Germany and the United Kingdom collaborate with
China directly without other European countries being involved, the new EU member
states—that have a relatively low scientific capacity—intend to cooperate with China
jointly with old EU member(s). The dense connections between the new EU members
(with middle- or low-capacity) reveal the emergence of research economies in Europe.
In the globalization era, scientific collaborations between China and the EU member states
have been greatly strengthened. To further facilitate and benefit from knowledge flows, it is
of importance to understand the structure and dynamics of collaboration networks. What is
more important for the EU side, the intra-European collaboration appears to be embedded
within the network of collaborations with China. Based on co-authorship data between
2000 and 2014 this paper examines the network structure of collaborations between China
and the EU 28, as well as the interactions between EU member states.
Our results show that the new China-EU collaboration network has been reshaped in
particular by a number of new EU member states that have middle or low research
capacities. These emerging research economies, as indicated by
changing the global balance of research activity. There are several reasons
behind this phenomenon. On one hand, collaborations between China and ‘old’ EU member
states—such as Germany, the UK and France—have been established for a relatively long
time, and the international partnership stays rather stable. However, there is not much
increasing space as the scientists in such large countries have a relatively higher possibility
to collaborate with domestic partners
. On the other hand, there is much
potential for collaboration between China and the middle- and low capacity nations as their
research is more internationally oriented. More and more Chinese researchers have come to
realize that there are also many excellent institutes and researchers besides those in the
large countries and have begun to seek collaborators from a wider range of regions,
including those relatively small countries. If the network cohesion represents a positive
relationship with the extent of information exchange
(Fritsch and Kauffeld-Monz 2010)
the majority of the new EU member states have been embedded in a better network
structure after they joined the EU.
What’s more, this study provides interesting evidence on the dynamic change of
collaborations between the old and new EU member states. In the later years, after joining the
EU, the new EU members have presented intensive collaborations with some pre-existing
EU members as well as strong interactions among the new members themselves. This
phenomenon can be partly explained by the argument of
Hoekman et al. (2013)
, that the
EU funding has a significant effect on scientific co-publications between member states
that did not intensively co-publish before, and in particular when involving scientifically
weak regions. This is also one of the goals of the European Research Area (ERA) concept,
which is to increase collaborations between European regions and contribute to
geographically integrated European research systems
(Scherngell and Lata 2013)
Using a dataset of co-publications between European regions in the period 2000–2007,
Hoekman et al. (2010)
find a gradual convergence toward a more integrated European
also points out that small and peripheral countries have
been more and more involved in European collaboration. Our results show that, in the
process of collaborating with China, such ‘‘Europeanisation’’ phenomenon is even more
obvious in recent years
after the participation of the new EU member
states. Small countries that joined the EU after 2000 have in particular enormously
improved their interconnection with other member states.
Another interesting observation from this study is that a few old EU member states play
a crucial role in forming the collaboration network.
Maes and Verdun (2005)
found that, in
addition to large EU member states, small countries such as Belgium and the Netherland
have played a significant role in the EU, e.g. in the creation of the economic and monetary
union in Europe. Our results demonstrate that, in terms of producing scientific output,
countries such as Belgium, Sweden, Denmark and the Netherlands, act as important
brokers connecting China and the new EU members.
Evidence shows that the international collaboration propensity in small and peripheral
European countries has increased enormously. Following this trend and with the rise of
China as a science powerhouse, it is expected that there will be more and more
collaborations between China and small European countries. Although having no advanced
research resources to provide, such small and peripheral countries can provide
complementarities to their partners. Nevertheless, economic distance (in terms of R&D capability)
is often a factor impeding core-peripheral collaborations
(Acosta et al. 2011;
Autantbernard et al. 2007)
. In such a case, the middle-level partners, who are open to and in favor
of European integration
(Maes and Verdun 2005)
, serve as gatekeepers that play an
important role in bridging China and the small partners in the EU.
One limitation of this study is that the networks are mapped based on full count, using
one type of collaboration measure (i.e. the Jaccard index). Although the Jaccard index is a
commonly used method in analysing co-author networks
(Pislyakov and Shukshina 2014;
, one should bear in mind that there are also other indicators, including
fractional values, Salton’s indexes and probabilistic measures (Probabilistic Affinity Index
and Probabilistic Partnership Index). Comparison between the results employing the
probabilistic measures and the Jaccard index is worthy of additional investigation.
Acknowledgements The authors would like to thank the anonymous referees for their helpful comments
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution,
and reproduction in any medium, provided you give appropriate credit to the original author(s) and the
source, provide a link to the Creative Commons license, and indicate if changes were made.
Abbasi , A. , Hossain , L. , & Leydesdorff , L. ( 2012 ). Betweenness centrality as a driver of preferential attachment in the evolution of research collaboration networks . Journal of Informetrics , 6 ( 3 ), 403 - 412 . doi: 10 .1016/j.joi. 2012 . 01 .002.
Acosta , M. , Coronado , D. , Ferra´ndiz, E., & Leo´n, M. D. ( 2011 ). Factors affecting inter-regional academic scientific collaboration within Europe: The role of economic distance . Scientometrics , 87 ( 1 ), 63 - 74 . doi: 10 .1007/s11192-010-0305-6.
Adams , J. ( 2012 ). The rise of research networks . Nature , 490 , 335 - 336 .
Autant-bernard , C. , Billand , P. , Frachisse , D. , & Massard , N. ( 2007 ). Social distance versus spatial distance in R&D cooperation: Empirical evidence from European collaboration choices in micro and nanotechnologies . Papers in Regional Science , 86 ( 3 ), 495 - 519 . doi: 10 .1111/j.1435- 5957 . 2007 . 00132 .x.
Beaver , D. de B , & Rosen , R. ( 1978 ). Studies in scientific collaboration . Scientometrics , 1 ( 1 ), 65 - 84 . doi: 10 . 1007/BF02016840.
Boschma , R. , Heimeriks , G. , & Balland , P. A. ( 2014 ). Scientific knowledge dynamics and relatedness in biotech cities . Research Policy , 43 ( 1 ), 107 - 114 . doi: 10 .1016/j.respol. 2013 . 07 .009.
Bound , K. , Saunders , T. , Wilsdon , J. , & Adams , J. ( 2013 ). China's Absorptive State: Research, innovation and the prospects for China-UK collaboration . London: Nesta.
Breschi , S. , & Lissoni , F. ( 2003 ). Mobility and social networks: localised knowledge spillovers revisited . CESPRI Working Paper, No. 142 , March 2003 .
Coccia , M. , & Wang , L. ( 2016 ). Evolution and convergence of the patterns of international scientific collaboration . Proceedings of the National Academy of Sciences , 113 ( 8 ), 2057 - 2061 . doi: 10 .1073/ pnas.1510820113.
Craig , P. , & De Bu´rca, G. ( 2015 ). EU law : text, cases, and materials (6th ed.). Oxford: Oxford University Press.
European Commission . ( 1995 ). A long term policy for China-Europe relations (Vol. July, COM).
European Commission . ( 2000 ). Communication from the Commission to the Council, the European Parliament, the Economic and Social Committee and the Committee of the Regions: Towards a European Research Area . COM ( 2000 ) 6 .
European Commission . ( 2014 ). European Research Area: Progress report 2014. doi:10 .2777/3450.
European Commission . ( 2016 ). Roadmap for EU-China S&T cooperation . Brussels, October 2016 .
European Council . ( 2000 ). Presidency conclusions . http://www.europarl.europa.eu/summits/lis1_en.htm ( 26 /12/ 2015 ).
Freeman , L. C. ( 1978 ). Centrality in social networks conceptual clarification . Social Networks , 1 ( 3 ), 215 - 239 . doi: 10 .1016/ 0378 - 8733 ( 78 ) 90021 - 7 .
Frenken , K. ( 2002 ). Europeanisation of science . Tijdschrift voor Economische en Sociale Geografie , 93 ( 5 ), 563 - 570 .
Fritsch , M. , & Kauffeld-Monz , M. ( 2010 ). The impact of network structure on knowledge transfer: an application of social network analysis in the context of regional innovation networks . The Annals of Regional Science , 44 ( 1 ), 21 - 38 . doi: 10 .1007/s00168-008-0245-8.
Gao , X. , Guan , J. , & Rousseau , R. ( 2011 ). Mapping collaborative knowledge production in China using patent co-inventorships . Scientometrics , 88 ( 2 ), 343 - 362 . doi: 10 .1007/s11192-011-0404-z.
Gla ¨nzel, W. ( 2001 ). National characteristics in international scientific co-authorship relations . Scientometrics , 51 ( 1 ), 69 - 115 . http://link.springer.com/content/pdf/10.1023/A:1010512628145.pdf.
Guan , J. , Zhang, J. , & Yan , Y. ( 2015 ). The impact of multilevel networks on innovation . Research Policy , 44 ( 3 ), 545 - 559 . doi: 10 .1016/j.respol. 2014 . 12 .007.
Hand , E. ( 2010 ). ''Big science'' spurs collaborative trend . Nature , 463 (January), 282 . doi: 10 .1038/463282a.
Hazir , C. S. , & Autant-Bernard , C. ( 2014 ). Determinants of cross-regional R&D collaboration: some empirical evidence from Europe in biotechnology . Annals of Regional Science , 53 ( 2 ), 369 - 393 . doi: 10 . 1007/s00168-014-0606-4.
Hoekman , J. , Frenken , K. , & Tijssen , R. J. W. ( 2010 ). Research collaboration at a distance: Changing spatial patterns of scientific collaboration within Europe . Research Policy , 39 ( 5 ), 662 - 673 . doi: 10 .1016/j. respol. 2010 . 01 .012.
Hoekman , J. , Frenken , K. , & van Oort , F. ( 2009 ). The geography of collaborative knowledge production in Europe . Annals of Regional Science , 43 , 721 - 738 . doi: 10 .1007/s00168-008-0252-9.
Hoekman , J. , Scherngell , T. , Frenken , K. , & Tijssen , R. ( 2013 ). Acquisition of European research funds and its effect on international scientific collaboration . Journal of Economic Geography , 13 ( 1 ), 23 - 52 . doi: 10 .1093/jeg/lbs011.
Katz , J. S. , & Hicks , D. ( 1997 ). How much is a collaboration worth? A calibrated bibliometric model . Scientometrics , 40 ( 3 ), 541 - 554 . doi: 10 .1007/BF02459299.
Katz , J. S. , & Martin , B. R. ( 1997 ). What is research collaboration? Research Policy, 26 ( 1 ), 1 - 18 . doi: 10 . 1016/S0048- 7333 ( 96 ) 00917 - 1 .
Leydesdorff , L. ( 2008 ). On the normalization and visualization of author co-citation data: Salton's cosine versus the Jaccard index . Journal of the American Society for Information Science and Technology , 59 ( 1 ), 77 - 85 . doi: 10 .1002/asi.20732.
Leydesdorff , L. , & Wagner , C. ( 2008 ). International collaboration in science and the formation of a core group . Journal of Informetrics , 2 ( 4 ), 317 - 325 . doi: 10 .1016/j.joi. 2008 . 07 .003.
Li , A. , & Chang , C. ( 2014 ). Beyond competition: Past, present and future on EU-China science and technology collaboration . European Foreign Affairs Review , 19 ( 3 ), 97 - 117 .
Luukkonen , T. , Tijssen , R. J. W. , Persson , O. , & Sivertsen , G. ( 1993 ). the measurement of international scientific collaboration . Scientometrics , 28 ( 1 ), 15 - 36 . doi: 10 .1007/BF02016282.
Maes , I. , & Verdun , A. ( 2005 ). Small states and the creation of EMU: Belgium and the Netherlands, pacesetters and gate-keepers . Journal of Common Market Studies , 43 ( 2 ), 327 - 348 . doi: 10 .1111/j.0021- 9886 . 2005 . 00558 .x.
Makkonen , T. , & Mitze , T. ( 2016 ). Scientific collaboration between ''old'' and ''new'' member states: Did joining the European Union make a difference? Scientometrics , 106 ( 3 ), 1193 - 1215 . doi: 10 .1007/ s11192-015-1824-y.
Mattsson , P. , Laget , P. , Nilsson , A. , & Sundberg , C. J. ( 2008 ). Intra-EU vs. extra-EU scientific co-publication patterns in EU . Scientometrics, 75 ( 3 ), 555 - 574 . doi: 10 .1007/sI1192-007-1793-x.
Melin , G. , & Persson , O. ( 1996 ). Studying research collaboration using co-authorships . Scientometrics , 36 ( 3 ), 363 - 377 . doi: 10 .1007/BF02129600.
Moed , H. F. ( 2016 ). Iran's scientific dominance and the emergence of South-East Asian countries as scientific collaborators in the Persian Gulf Region . Scientometrics, 108 , 305 - 314 . doi: 10 .1007/s11192- 016-1946-x.
Morescalchi , A. , Pammolli , F. , Penner , O. , Petersen , A. M. , & Riccaboni , M. ( 2015 ). The evolution of networks of innovators within and across borders: Evidence from patent data . Research Policy , 44 ( 3 ), 651 - 668 . doi: 10 .1016/j.respol. 2014 . 10 .015.
Narin , F. , Stevens , K. , & Whitlow , E. S. ( 1991 ). Scientific co-operation in Europe and the citation of multinationally authored papers . Scientometrics , 21 ( 3 ), 313 - 323 . doi: 10 .1007/BF02093973.
Newman , M. E. J. ( 2004 ). Coauthorship networks and patterns of scientific collaboration . Proceedings of the National Academy of Sciences, 101(Supplement 1) , 5200 - 5205 .
Nomaler , O¨ . , Frenken , K. , & Heimeriks , G. ( 2013 ). Do more distant collaborations have more citation impact ? Journal of Informetrics , 7 ( 4 ), 966 - 971 . doi: 10 .1016/j.joi. 2013 . 10 .001.
Pislyakov , V. , & Shukshina , E. ( 2014 ). Measuring excellence in Russia: Highly cited papers, leading institutions, patterns of national and international collaboration . Journal of the Association for Information Science and Technology , 65 ( 11 ), 2321 - 2330 . doi: 10 .1002/asi.
Price , D. J. de S., & Beaver , D. de B. ( 1966 ). Collaboration in an invisible college . American Psychologist , 21 ( 11 ), 1011 - 1018 . doi: 10 .1037/h0024051.
Revilla , E. , Sarkis , J. , & Modrego , A. ( 2003 ). Evaluating performance of public-private research collaborations: A DEA analysis . Journal of the Operational Research Society , 54 ( 2 ), 165 - 174 . doi: 10 .1057/ palgrave.jors. 2601524 .
Scherngell , T. , & Barber , M. J. ( 2011 ). Distinct spatial characteristics of industrial and public research collaborations: Evidence from the fifth EU Framework Programme . Annals of Regional Science , 46 ( 2 ), 247 - 266 . doi: 10 .1007/s00168-009-0334-3.
Scherngell , T. , & Lata , R. ( 2013 ). Towards an integrated European research area? Findings from Eigenvector spatially filtered spatial interaction models using European framework programme data . Papers in Regional Science , 92 ( 3 ), 555 - 577 . doi: 10 .1111/j.1435- 5957 . 2012 . 00419 .x.
Suttmeier , R. P. ( 2014 ). Trends in U.S. -China Science and Technology Cooperation: Collaborative Knowledge Production for the Twenty-First Century? By Research Report Prepared on Behalf of the U.S.-China Economic and Security Review Commission . http://search.usa.gov/search?query= research&op=Search&affiliate=uscc.gov.
Tang , L. ( 2013 ). Does ''birds of a feather flock together'' matter-Evidence from a longitudinal study on USChina scientific collaboration . Journal of Informetrics , 7 ( 2 ), 330 - 344 . doi: 10 .1016/j.joi. 2012 . 11 .010.
Tang , L. , & Shapira , P. ( 2011 ). China-US scientific collaboration in nanotechnology: patterns and dynamics . Scientometrics , 88 ( 1 ), 1 - 16 . doi: 10 .1007/s11192-011-0376-z.
Tijssen , R. J. W. ( 2008 ). Are we moving towards an integrated European Research Area? Collnet Journal of Scientometrics and Information Management , 2 ( 1 ), 19 - 25 . doi: 10 .1080/09737766. 2008 . 10700837 .
Toivanen , H. , & Ponomariov , B. ( 2011 ). African regional innovation systems: Bibliometric analysis of research collaboration patterns 2005-2009 . Scientometrics, 88 ( 2 ), 471 - 493 . doi: 10 .1007/s11192-011- 0390-1.
Wagner , C. , & Leydesdorff , L. ( 2005 ). Network structure, self-organization, and the growth of international collaboration in science . Research Policy , 34 ( 10 ), 1608 - 1618 . doi: 10 .1016/j.respol. 2005 . 08 .002.
Wang , L. ( 2016 ). The structure and comparative advantages of China's scientific research: quantitative and qualitative perspectives . Scientometrics , 106 ( 1 ), 435 - 452 . doi: 10 .1007/s11192-015-1650-2.
Wang , L. , & Wang , X. ( 2017 ). Who sets up the bridge? Tracking scientific collaborations between China and the European Union . Research Evaluation, 10 , 10. doi: 10 .1093/reseval/rvx009.
Wang , X. , Xu , S. , Liu , D. , & Liang , Y. ( 2012 ). The role of Chinese-American scientists in China-US scientific collaboration: A study in nanotechnology . Scientometrics , 91 ( 3 ), 737 - 749 . doi: 10 .1007/ s11192-012-0693-x.
Wang , X. , Xu , S. , Wang , Z. , Peng , L. , & Wang , C. ( 2013 ). International scientific collaboration of China: Collaborating countries, institutions and individuals . Scientometrics , 95 ( 3 ), 885 - 894 . doi: 10 .1007/ s11192-012-0877-4.
Yamashita , Y. , & Okubo , Y. ( 2006 ). Patterns of scientific collaboration between Japan and France: Intersectoral analysis using Probabilistic Partnership Index (PPI) . Scientometrics , 68 ( 2 ), 303 - 324 . doi: 10 . 1007/s11192-006-0105-1.
Yan , E. , & Guns , R. ( 2014 ). Predicting and recommending collaborations: An author- , institution- , and country-level analysis . Journal of Informetrics , 8 ( 2 ), 295 - 309 . doi: 10 .1016/j.joi. 2014 . 01 .008.
Zhou , P. , & Gla¨nzel, W. ( 2010 ). In-depth analysis on China's international cooperation in science . Scientometrics , 82 ( 3 ), 597 - 612 . doi: 10 .1007/s11192-010-0174-z.
Zhou , P. , Zhong , Y. , & Yu , M. ( 2013 ). A bibliometric investigation on China-UK collaboration in food and agriculture . Scientometrics , 97 ( 2 ), 267 - 285 . doi: 10 .1007/s11192-012-0947-7.
Zitt , M. , & Bassecoulard , E. ( 2004 ). S& T networks and bibliometrics: The case of international scientific collaboration . In Fourth Congress on Proximity Economics: Proximity, Networks and Coordination (pp. 17 - 18 ).
Zitt , M. , Bassecoulard , E. , & Okubo , Y. ( 2000 ). Shadows of the past in international cooperation-collaboration profiles of the top five producers of science . Scientometrics , 47 ( 3 ), 627 - 657 .