Scholarly Collaboration between Europe and Israel:
A Scientometric Examination of a Changing Landscape
Eric Zimmerman[1]
Academic Secretary and Director of theResearchAuthority
InterdisciplinaryCenter Herzliya (IDC)
P.O. Box 167Herzliya46150Israel
Tel: +972.9.952.7676 Fax: +972.9.952.7268
Wolfgang Glänzel
KU Leuven
Steunpunt O&O Statistieken
Dekenstraat 2, B-3000 Leuven, Belgium
and
Judit Bar-Ilan
Department of Information Science, Bar-IlanUniversity
Ramat Gan 52900 Israel
Tel: +972.3.5318351 Fax: +972.3.7384027
Abstract
In this paper we examine various aspects of the scientific collaboration between Europe and Israel, and show that the traditional collaboration patterns of Israel (preference towards collaboration with the US) is changing, and the collaboration with the EU countries is growing.
Introduction
The international flavor of Israeli research and development is changing. Until recently the United States has been the major foreign supporter of sciencein Israel. Recently, however, other international players have entered the field - notably the European Union through the Research Framework Programme,[2]is altering the Israeli S&T landscape – creating a new alliance.
There is evidence of a decrease inUSA federal funding levels vis-à-vis Israeli basic research. Some argue that this funding decrease, combined with Israeli's entry into the European Union Research Framework Programme signals a shift in Israel's orientation away from The United States. Empirical evidence from the National Science Foundation (NSF) and Science Citation Index Expanded (SCIE) based on co-authorship trends support this assertion (Glänzel et al., 2007b). Europe will soon reach parity (if not surpass) the USAboth in the funding of research and in scientific collaborationwithIsrael.
With regard to article authorship statistics, the percent of papers published by Israeli researchers with non-Israeli partners which are co-authored with US researchers has indeed gone down. According to the calculations prepared for The United States-Israel Educational Foundation (USIEF) by Gideon Czapski of the HebrewUniversity in Jerusalem (HUJI)[3] on the basis of the SCI data base, US coauthored articles accounted for 68.2% of all internationally coauthored articles in 1986, but only 53.6% in 1999.[4]Our observations have confirmed the continuation of this trend for the new millennium, as well. Additionally, there has been a decline in the overall percent of the total world output of scientific articles with US involvement. According to a recent study onglobal changes in the scientific landscape of the world (Glänzel et al., 2008)while US participation was indicated in 35.6% of all articles published in 1991, it declined to 30.4% in 2005.
The decline in the relative prominence of the US as a scientific partner for Israel as measured by article co-authorship may be linked to a decline in the relative prominence of the US in world science overall (Leydesdorff and Wagner, 2007). Taking into account both that the EU15 also is losing its share of the world total since around 2000, mirroring the relative decline of the USA and Japan (Glänzel et al., 2007a), and that the intensification of EU-Israel co-operation has continued after 1998, this global trend cannot explain all aspects of the evolution of bilateral relations. Thus, the ‘weight’ of the US might be lessening because the ‘weight’ of others like China, Brazil, Taiwan, Korea, India and Turkey is rising,but this phenomenon presently holds to a lesser extent for the EU, too (Zhou and Leydesdorff, 2006, Glänzel et al., 2008). The drop in the U.S. share of international co-authored journal publications is a well-known, long-term phenomenon which has recently been studied by Glänzel et al. (2007a).
The emergence of the European Framework Programmes has resulted in greater cooperation between European countries among themselves and with others. This is true for Israel, Greece, and other countries that had been focused on the US but have since widened their cooperation net, and this has included increases within Europe. The increase in collaboration has been documented in recent European studies (e.g., REIST-2, 1977; REIST-3, 2003).
In the present study we summarize recentresearch, describingthe shifting orientation of Israeli science towards Europe – measured by increasing funding opportunities and grant amounts, growing co-authorships, rising numbers of technology transfer agreements, and patents registered in Europe. For this analysis we will use data from varied sources, including the Office of the Chief Scientist and longitudinal data sets, spanning two decades, and refer to the observations made by Glänzel et al. (2007b).
We alsoanalyze the co-authorship network of Israel in 1991 and 2005, anddiscuss the Israeli culture/model of research and innovation financing (e.g. entrepreneurship, venture capital and Office of the Chief Scientist programs) and how this might serve as a model for Europe, given the Lisbon Agenda aimed at promoting economic growth, fostering competitiveness and stimulating job creation. The Lisbon Agenda foresaw European countries spending 3% of their GDP on R&D, by 2010. This will not be reached. Already, for some time, however, Israel is placed above the 4% mark.
Data sources
Most of the publication-related data are based on bibliographic data extracted from the annual volumes of the Science Citation Index Expanded (SCIE) of Thomson Scientific (Philadelphia, PA, USA). The Science Citation Index Expanded, whichis part of the Web of Science, indexes approximately 6000 of the world's leading scientific journals in the life sciences, natural sciences, mathematics and engineering. The multidisciplinary database has unique features that allow its use for bibliometric purposes.
We have only original research documents, that is, document types named as articles, letters, notes and reviews taken into consideration. Publications were assigned to countries on the basis of their corporate addresses which appear in the by-line of the publication. All countries indicated in the address field were considered thus allowing for the analysis of international co-authorship as well. Citations received by these papers have been counted for a three-year citation window beginning with the publication year on the basis of an item-by-item procedure using special identification-keys made up of bibliographic components.
As for subject classification, publications were arranged into 12 major fields: Agriculture & Environment (AGRI), Biology (BIOL), Biosciences (BIOS), Biomedical research (BIOM), Clinical & Experimental Medicine I – General & Internal Medicine (CLI1), Clinical & Experimental Medicine II – Non-Internal Medicine Specialties (CLI2), Neuroscience & Behaviour (NEUR), Chemistry (CHEM), Physics (PHYS), Geosciences & Space Sciences (GEOS), Engineering (ENGN) and Mathematics (MATH). This classification scheme was developed by GlänzelandSchubert(2003) on the basis of the ISI Subject Categories for journals.
Methods and Results
The Israeli R&D Climate
Small in size and population, lacking many natural resources and burdened with imposing defense needs, Israel has nonetheless come a long way since the establishment of the State in 1948. Since 1950 Israel has witnessed an almost fivefold growth in its population and in its per capita GDP in real terms. No explanation of this economic growth would be complete without taking into account Israel’s strong, vibrant and highly dynamic research system. Israel maintains a long-established record of activity and achievements in basic and applied scientific research.
The following are some indicators of Israel’s strength in science:
- In recent years three Israeli scientists have been awarded the Nobel Prize
- Israel ranks 17th in terms of the number of citations and 21st in terms of the number of published papers among the 145 countries/territories monitored by the ISI’s Essential Science Indicators (Thomson, 2007) and based on publications in the last ten years.
- In relative terms, Israel ranks third in the world, behind Switzerland and Sweden, in the number of scientific articles per capita and seventh in the world in citations per capita (de Fontenay and Carmel, 2001).
- Israel ranks among the top ten countries in terms of citations, based on the last tens years of publications in computer science, psychiatry/psychology and social sciences, general and its rank is between 11 and 16 (i.e., higher than its overall rank) in biology & biochemistry, immunology, mathematics, molecular biology & genetics and in neuroscience & behavior (Thomson, 2007). Here the data is for the ISI subject classification into twenty two fields and not into the twelve major fields used in other parts of the study.
- R&D intensity in Israel in 2001 represents 4.8% of GDP - of the highest ratios alongside 4.27% in Sweden, 3.49 inFinland, 3.06% in Japan and 2.8% in the US (European Commission, 2003).
- Leading international academic institutions - national science citation intensity, measured as the ratio of scientific citations of all papers to the national GDP places Israel alongside the strongest performers by this measure: Switzerland, Sweden, Finland and Denmark (Cordis, 2007)
Israel’s research is a reflection of its scientific publications
According to the Science Citation Index Expanded (SCIE), Israel adds about 1.2% of the all publications to the world total (see Figure 1). This contribution is (except for some annual fluctuations) quitestable. Israel’s share in the world total lies actually between that of Denmark and Belgium (cf. Glänzel et al., 2007a). However, concerning the publication output per capita Israel ranks at the top immediately behind Sweden and Switzerland. The cross-national comparison of this indicator is presented in Figure 2.
Source: Science Citation Index Expanded (Thomson-Scientific)
Figure 1: Scientific publications of Israel as percentage of the World’s publication output
Source: Science & Engineering Indicators 2006, US NSF
Figure 2: Scientific Publications - 2000-2003 and Number of Citations in Scientific Literature - 2003 per Million Population (USA = 1.00)
Although scientific collaboration cannot always be depicted by co-authorship in an adequate manner (e.g., Katz and Martin, 1997), joint publicationsare one of the most tangible and well documented forms of collaboration in research.Nonetheless, international co-publicationsproved a good indicator of co-operation at this level (Glänzel and Schubert, 2004).
A first look at the publication data (Glänzel et al., 2007b) reveals a strong increase of Israel’s international level of cooperation. Israel increased publication output according to the SCIE by about two thirds in the period 1991-2005 but the number of internationally co-authored papers has more than doubled in the same time. Thus, Israel’s share of internationally co-authored papers in the sciences grew from 31.8% in 1991 to 38.8% in 1998 and finally to 41.9% in 2005. With regard to authorship statistics, the percent of papers published by Israeli researchers and co-authored with US authors, that is, the share of US-Israeli co-operation in all internationally co-authored publications of Israelhas indeed gone down. The share of EU15 collaboration in all Israeli ‘international’ papers, on the other hand, has considerably grown. This trend is presented in Figure 3. Nevertheless, with more than 50% of all international papers of Israel, the United States is still Israel’s most important partner. These results are in line with the calculations prepared for the United States-Israel Educational Foundation (USIEF) by Gideon Czapski[5](HebrewUniversity in Jerusalem, HUJI).
Source: Science Citation Index Expanded (Thomson-Scientific)
Figure 3.The evolution of the share of US and EU co-authorship in all Israeli
publications
A further look reveals that among the European countries Germany, France, UK, Italy and the Netherlands are presently Israel’s most important partners in Europe. The evolution of their contribution to Israel’s international co-publications can be found in Table 1. Only those countries, which are contributing with at least 5% to all international co-publications of Israel in 2005, are presented here.
Table 1. The share of international collaboration of Israel’s most important partners in all Israeli co-publications in 1991, 1998 and 2005
Country / 1991 / 1998 / 2005USA / 64.3% / 55.1% / 53.2%
Germany / 12.2% / 16.1% / 15.2%
France / 6.8% / 9.0% / 10.6%
UK / 6.5% / 9.4% / 9.5%
Italy / 2.4% / 5.9% / 7.4%
Canada / 6.0% / 6.8% / 6.7%
Netherlands / 1.8% / 4.0% / 5.3%
Source: Science Citation Index Expanded (Thomson-Scientific)
The question arises of in how far this trend mirrors the decline in the overall percentage of the US output in world total in scientific articles that has already been reported in several European studies (e.g., REIST-2, 1997; REIST-3, 2003). The share of the US output was shrinking from 35.6% in 1991 to 30.5% in 2005 (cf. Glänzel et al., 2008). The decline in the relative prominence of the US as a scientific partner for Israel as measured by article co-authorship may to a certain extent be explained by a decline in the relative prominence of the US in the world science overall. Taking into account that also the EU15 is losing weight in the world total since about the Millennium change, mirroring the relative decline of the USA and Japan (Glänzel et al., 2008), and intensification of EU-Israel co-operation has continued after 1998, this global trend cannot explain all aspects of the evolution of bilateral relations alone. Thus, the ‘weight’ of the US might be lessening because the ‘weight’ of others like China, Brazil, Taiwan, Korea, India and Turkey is rising,but this phenomenon presently holds to a lesser extent for the EU15, too (Zhou and Leydesdorff, 2006, Glänzel et al., 2008). Although changes in national publication output might not have an immediate effect on bilateral relations, they do influence the strength of bilateral co-publication links.
We use Salton’s (cosine) measure as an indicator of international collaboration strength. This measure is defined as the number of joint publications divided by the square root of the product of the number (i.e., the geometric mean) of total publication outputs of the corresponding pair of countries (cf. Glänzel, 2001). Consequently, the strength of a bilateral co-operation might change even if the share of bilateral papers in the output of one of the countries is unchanged but that of the other one increases or decreases. Thus, Figure 4 supplements Figure 3on the evolution of the share of US and EU co-authorship by taking into account the publication dynamics of Israel’s partners, as well. According to the Salton measure, the strength of co-operation link with the US is still increasing, however, to a lesser extent than that with the EU. The change of scholarly co-operation between Israel and Europe can best be visualized by ‘scientopograhical’ maps. Figure 4 presents Israel’s most important scientific co-operation partners in the world in 1991 and 2005. We have used three different thresholds1.0%, 2.5% and 5.0% to visualise the intensity of co-operation. The strength of 5% is however not reached by any link. The growing number of medium strong links substantiates the increasing role of Europe as Israel’s partner in scientific research.
Source: Science Citation Index Expanded (Thomson-Scientific)
Figure 4.Co-authorship map for Israel in all fields combined in 1991(left) and 2005 (right) based on Salton’s measure (dotted line ≥ 1.0%, solid line ≥ 2.5%, thick line ≥ 5.0%)
In the following we will have a look at the profile and the impact of collaborative research in Israel.International co-authorship is expected to result, on an average, in publications with higher citation rates than purely domestic papers. However, the influence of international collaboration on the national citation impact generally varies considerably between the countries (and within one individual country between fields) but in some cases there might be even no citation advantage for one or even for both partners (cf. Glänzel and Schubert, 2001). In order to analyze the citation impact of Israel’s bilateral publication links we have used the following indicators.
i)Activity Index (AI) is defined as the ratio of the share of a given field in the publications of a given country to the share of the same field in the world total publications. This indicator was originally used in economics as Comparative Advantage Index typically calculated with export data; it has later been introduced by Frame (1977) in bibliometrics, and is used in macro studies (e.g., Schubert et al., 1989). AI takes its values in the range [0, ], where its neutral value is 1. AI = 0 indicates a completely idle research field, AI < 1 indicates a lower-than-average and AI > 1 a higher-than-average activity. AI reflects the internal balance among the fields in the given country, that is, AI > 1 values in several fields must always be balanced by AI < 1 in others: AI cannot always be greater (less) than 1. Here a version of the Activity Index is used to compare the subject profile of Israel’s internationally co-authored papers with that of all Israeli publications.
ii)Mean Observed Citation Rate (MOCR). MOCR is defined as the ratio of citation count to publication count. It reflects the factual citation impact of a country, region, institution, research group etc. A three-year citation window has been applied.
iii)Mean Expected Citation Rate (MECR). The expected citation rate of a single paper is defined as the average citation rate of all papers published in the same journal in the same year. Instead of the one-year citation window to publications of the two preceding years as used in the Journal Citation Report (JCR), a three-year citation window to one source year is used, as explained above. For a set of papers assigned to a given country, region or institution in a given field or subfield, the indicator is the average of the individual expected citation rates over the whole set.
iv)Relative Citation Rate (RCR). RCR is defined as the ratio of the Citation Rate per Publication to the Expected Citation Rate per Publication, that is, RCR = MOCR/MECR. This indicator measures whether the publications of a country or institution attract more or less citations than expected on the basis of the impact measures, i.e., the average citation rates of the journals in which they appeared. Since the citation rates of the papers are gauged against the standards set by the specific journals, it is largely insensitive to the big differences between the citation practices of the different science fields and subfields. It should be stressed that in this study, a 3-year citation window to one source year is used for the calculation of both the enumerator and denominator of RCR. RCR = 0 corresponds to un-citedness, RCR < 1 means lower-than-average, RCR > 1 higher-than-average citation rate, RCR = 1 if the set of papers in question attracts just the number of citations expected on the basis of the average citation rate of the publishing journals. RCR has been introduced by Schubert et al. (1983), and largely been applied to comparative macro and meso studies since. It should be mentioned that a version of this relative measure, namely, CPP/JCSm is used at CWTS in Leiden (see Moed et al, 1995).
v)Normalised Mean Citation Rate (NMCR). NMCR is defined analogously to the RCR as the ratio of the Mean Observed Citation Rate to the weighted average of the mean citation rates of subfields. This indicator is a second expected citation rate; in contrast to the RCR, NMCR gauges citation rates of the papers against the standards set by the specific subfields. Its neutral value is 1 and NMCR >(<) 1 indicates higher(lower)-than-average citation rate than expected on the basis of the average citation rate of the subfield. NMCR has been introduced by Braun and Glänzel(1990) in the context of national publication strategy. A similar measure (CPP/FCSm) is used at CWTS (cf. Moedet al, 1995).