This study presents a preliminary comparison of networks of international collaboration and mobility. Using affiliation data from scientific publications, we analyse the structural differences in the two networks and the role of countries. The results show that researchers collaborate internationally to a much higher degree than they become internationally mobile. The number of countries involved in the networks is three times higher in collaboration than in mobility, and the average degree demonstrates that mobility networks form tight structures with fewer links than collaboration networks. The role of countries differs between the collaboration and mobility network, predominately reflecting income level. Limitations and future research are described to further understand the dynamics of collaboration and mobility networks.

1. Networks of international collaboration and
mobility: a comparative study
ISSI 2017, 16-20 October, Wuhan, China
Zaida Chinchilla-Rodríguez, Lili Miao, Dakota Murray, Nicolás Robinson-
García, Rodrigo Costas and Cassidy R. Sugimoto

2. Dimensions of conectivity
International collaboration
Scientific mobility
ISSI Conference 2017 2

3. Dimensions of conectivity
International collaboration plays an important role
in fostering high-quality knowledge production
ISSI Conference 2017 3

4. Dimensions of Connectivity
• Mobility…serves to both
strengthen the
scientific capacity of
the countries and
benefit the scientific
careers of individual
researchers (OECD
2008)
ISSI Conference 2017 4

5. Dimensions of Connectivity
ISSI Conference 2017 5
…mobility has been advocated as key to increasing the efficiency
and effectiveness of research (Van Noorden, 2012: Scellato et al.
2015)

6. Immigration ban
Some political leaders
are proposing and
implementing
nationalistic policies
that restrict global
mobility
ISSI Conference 2017 6
+ North Korea
+ Venezuela

7. Science knows no border
Limiting mobility is likely to
have adverse effects on the
scientific system, which is
increasingly dependent on
international collaboration
ISSI Conference 2017 7
Sugimoto, C.R., Robinson-García, N., Murray, D. R., Yegros-Yegros, A., Costas, R. & Lariviere, V.
(2017). Scientists have most impact when they are free to move. Nature, 32(550), 29-31

8. The idea
Comparing the two dimensions of connectivity
could improve the understanding of mobility
and identify opportunities for and risks of
scientific collaboration and mobility
ISSI Conference 2017 8

9. Our Goal
To understand whether mobility provides a
novel or a duplicative lens on the network
created through scientific collaboration
ISSI Conference 2017 9
Preliminary comparative analysis of networks
and role of more than 200 countries
considering economic and scientific capacities

10. ESTABLISH A TERMINOLOGY
ISSI Conference 2017 10

11. International Parternships
• Scientific collaboration
• Scientific mobility
– Co-affiliation
– Migration
ISSI Conference 2017 11

12. Collaboration
ISSI Conference 2017 12

13. Example: this paper!
ISSI Conference 2017 13

14. International co-affiliation
Holding more than one affiliation in two
countries during a period of study
ISSI Conference 2017 14

15. t0
ISSI Conference 2017 15

16. t1
ISSI Conference 2017 16

17. Linked by co-affiliation
ISSI Conference 2017 17

18. International Migration
More permanent movements between
affiliations in 2 or more countries
ISSI Conference 2017 18

19. t0
ISSI Conference 2017 19

20. t1
ISSI Conference 2017 20

21. t2
ISSI Conference 2017 21

22. Linked by migration
ISSI Conference 2017 22

23. Scientific Mobility
both co-affiliation and migration
ISSI Conference 2017 23

24. DATA
ISSI Conference 2017 24

25. Web of Science
ISSI Conference 2017 25
2008 – 2015
First publication in 2008 – 2015

26. Disambiguation
Caron and van Eck (2014)
John Smith
ISSI Conference 2017 26
Caron, E., & Van Eck, N.J. (2014). Large scale author name disambiguation using
rule-based scoring and clustering. In Proceedings of the 19th International
Conference on Science and Technology Indicators (pp. 79-86).

27. Final Dataset
ISSI Conference 2017 27
3,251,797 14,097,939 213 Countries

28. COUNTRY-LEVEL CLASSIFICATIONS
ISSI Conference 2017 28

29. Scientific capacity of nations
• Scientific Capacity is the
ability to use specialized
knowledge and exploit it
to conduct research
(Wagner et al 2001).
• Investment, infrastructure,
and output
• Four groups: advanced,
proficient, developing and
lagging
ISSI Conference 2017 29

30. Income level
ISSI Conference 2017 30(World Bank, 2016)

31. METHODS
ISSI Conference 2017 31

32. Total number of publications
ISSI Conference 2017 32
+8

33. Number of international publications
ISSI Conference 2017 33
+3+1

34. Total number of active researchers
ISSI Conference 2017 34
+5

35. Number of mobile researchers
ISSI Conference 2017 35
+1+2

36. Matrices of collaboration and mobility
• Co-ocurrences
frequencies
• Pajek 64 4.10,
NetworkX 1.10
• Undirected
• Unweighted
• Derive statistical
properties
ISSI Conference 2017 36
MobilityCollaboration

37. Structural indicators
• Density
• Average degree
• Average shortest distance
• Diameter
• Assortativity
• Closessneess
• Betweenness
• Clustering coefficient
ISSI Conference 2017 37

38. RESULTS (I)
Comparison of collaboration and mobility networks
ISSI Conference 2017 38

39. Summary of the properties of the collaboration and
mobility networks
Collaboration Mobility
Number of nodes 212 212
Number of edges 12596 4788
Density 0.55 0.21
Average degree 118.27 44.95
Average shortest
distance
1.44 1.84
Diameter 2 4
Cluster Coefficient 0.83 0.73
Assortativity -0.19 -0.34
~ 3
+

40. A summary of the properties of the collaboration and
mobility networks
Collaboration Mobility
Number of nodes 212 212
Number of edges 12596 4788
Density 0.55 0.21
Average degree 118.27 44.95
Average shortest
distance
1.44 1.84
Diameter 2 4
Cluster Coefficient 0.83 0.73
Assortativity -0.19 -0.34

41. RESULTS (II)
Comparison of the position in closeness, betweenness and
clustering coefficient in the four groups of countries according to
their scientific and economic capacities
ISSI Conference 2017 41

42. How to interpret the following section
• For each indicator:
collaboration and
mobility
• The colours of lines
refer to the income
level: (green: high,
blue: upper-middle,
orange: low-middle
and red: low)

43. Closeness among the scientifically advanced countries
• Netherlands (8) (9): it
ranks in the eighth
position in the
collaboration network
and in the ninth
position in the mobility
network
• Lines only show
variations into the group

44. Scientifically advanced countries
ISSI Conference 2017 44
Clossenes Betweenness Clustering coefficient

45. Scientifically advanced countries
ISSI Conference 2017 45
Clossenes Betweenness Clustering coefficient

46. Scientifically advanced countries
ISSI Conference 2017 46
Clossenes Betweenness Clustering coefficient

47. Scientifically proficient countries
Clossenes Betweenness Clustering coefficient

48. Scientifically developing countries
Legend: Legend: The colours refer to the income level: (green: high, blue: upper-middle,
orange: low-middle and red: low)
Clossenes Betweenness Clustering coefficient

49. Scientifically developing countries
Legend: Legend: The colours refer to the income level: (green: high, blue: upper-middle,
orange: low-middle and red: low)
Clossenes Betweenness Clustering coefficient

50. SUMMARY
ISSI Conference 2017 50

51. Preliminary large-scale analysis in
establishing international relationships
ISSI Conference 2017 51STI Conference 2017

52. International Partnership
Traditionally, large-scale analyses have examined
only collaboration or mobility
ISSI Conference 2017 52

53. Well-studied “elite” countries
ISSI Conference 2017 53

54. Understudied Periphery
ISSI Conference 2017 54

55. Moving away from the elite
To understand the global
system of science, we
need to move away from
the “elite” and towards a
comprehensive analysis of
international scientific
partnership
ISSI Conference 2017 55

56. Deviance from assumptions
• An small subset of countries such as India,
Malaysia, Turkey and Saudi Arabia builds
stronger connections in the mobility network.
• We also highlight the important role of South
Africa, Kenya and Nigeria in collaboration
network
ISSI Conference 2017 56

57. Collaboration related to mobility
• Both networks show a high level of local
clustering and a small average number of
steps between actors
ISSI Conference 2017 57

58. But not the same
• A country’s position is not always
consistent in both networks revealing
that roles of countries in terms of
capacities to be accessible to other
nodes, in transfer knowledge, or to be
cohesive with their neighbours may
differ.
ISSI Conference 2017 58

59. Scientific relationships are highly
resource-dependent
• Where estability
between networks
exists, it is mostly
associated wiht high-
income countries
ISSI Conference 2017 59

60. IMPLICATIONS
ISSI Conference 2017 60

61. ≠
ISSI Conference 2017 61STI Conference 2017 61
Not all partnerships are equal

62. Discrepancy
• Some countries have high importance in one
dimension and low in the other one
dimension
• The discrepancy between two dimensions
could be an indicator of potential
opportunities for scientific collaboration and
mobility programs
ISSI Conference 2017 62

63. Many perspectives needed
ISSI Conference 2017 63
• Evaluative
frameworks of
countries must
consider a variety
of indicators

64. LIMITATIONS AND FUTURE WORK
ISSI Conference 2017 64

65. Hetereogenity
• Preliminary, large-scale
analysis
• No consideration of
disciplinary context
• Positions in the worlwide
ranking
• Re-visiting classifications
ISSI Conference 2017 65

66. Future Approaches
Network analysis:
- new indicators
- new approaches
ISSI Conference 2017 66

67. Future Approaches
Network analysis
Time dimension
ISSI Conference 2017 67

68. Future Approaches
Network analysis
Time dimension
More and diverse
Indicators
ISSI Conference 2017 68

69. Future Approaches
Network analysis
Time dimension
Diverse Indicators
Mobility Flows
ISSI Conference 2017 69

70. Networks of international collaboration and
mobility: a comparative study
ISSI 2017, 16-20 October, Wuhan, China
Thanks for your attention!
Questions?
zaida.chinchilla@csic.es