Eight clusters : a dynamic perspective and structural analysis for the evaluation of institutional research performance
Thijs, B.
Citation
Thijs, B. (2010, January 27). Eight clusters : a dynamic perspective and structural analysis for the evaluation of institutional research performance. Retrieved from
https://hdl.handle.net/1887/14617
Version: Not Applicable (or Unknown)
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/14617
Note: To cite this publication please use the final published version (if applicable).
4 ISRAELI RESEARCH INSTITUTES: A
DYNAMIC
PERSPECTIVE
Bart Thijs1, Eric Zimmerman2, Judit Bar-Ilan3, Wolfgang Glänzel1,4
1Bart.Thijs@econ.kuleuven.be, Wolfgang.Glanzel@econ.kuleuven.be
Katholieke Universiteit Leuven, Steunpunt O&O Indicatoren, Dekenstraat 2, B-3000 Leuven, Belgium
2zimmee@idc.ac.il
Interdisciplinary Center Herzliya (IDC), Israel.
3barilaj@mail.biu.ac.il
Department of Information, Science, Bar-Ilan University, Ramat Gan, Israel
4glanzw@iif.hu
Hungarian Academy of Sciences, IRPS, Budapest, Hungary
Published in: Research Evaluation (2009), 18(3), 251-260 DOI: 10.3152/095820209X466900
The present paper follows up earlier studies of Israel’s changing research landscape. In those papers, among others, various aspects of scientifi c collaboration with Europe have been examined during a period of 15 years. The present study aims at analysing changing research profi les and performance of the fi ve important Israeli institutes from a dynamic perspective.
4.1 Introduction
The quality and reliability of tools used for the description of a research institute’s performance used by institutes, funding agencies and policy makers are increasingly questioned, as these tools are used in evaluation of research and funding decision-making for research institutes. The academic community faces increasing calls for transparency and accountability, and in the growing face of international ranking schemes, these tools are of increasing importance. Macro- level analysis is has been successfully used in bibliometrics to monitor national research performance from the perspective of international competitiveness, collaboration and globalisation. The identi cation of main actors and centres of excellence has recently become one of the favourite tasks in research evaluation (Leta et al, 2005, CWTS (Van Raan, 1999), ANU (Butler, 2007), OST etc.). Finding a common platform for the evaluation of research institutions with different pro les proved one of the greatest challenges in this endeavour. In this context we mention the university-ranking exercise (e.g., ARWU, 2007, CHE, 2007, THES- QS, 2007, Glänzel and Debackere, 2008) just as an example. A classi cation model for research institutes based on research activity in distinct scienti c elds has recently been developed by Thijs and Glänzel (2008). Scholarly publications of these institutes that are indexed by the ISI Citation Indexes are assigned to different scienti c elds. One of the main goals when creating this model was to improve the comparability of performance studies. Another application of this model is to examine the dynamics of research pro les and the growth of an individual institute.
4.2 The dynamic perspective
In the development of the methodology for our dynamic perspective we relied on four main principles
• National situation: Research institutes are embedded in a local science system with its own properties and peculiarities. They depend on local funding and are part of an established cultural heritage. This means that national statistics can provide a good starting point or reference point for the evaluation of institutional research performance. However, as institutes can have a much more specialized research pro le than the country, the national situation alone cannot be used as a normative benchmark.
• Evolution: A snapshot of an institute’s performance can give a clear idea of the relative position of this institute. However, changes in methodology, in the set of used indicators or in the underlying database, can strongly in uence a subsequent re-run of these snapshots making them incomparable. Therefore, we suggest the usage of evolution-data to reveal possible trends in the performance, and point out strengths and weaknesses.
• Multiple Indicators and Field differentiation: Research is a multi- faceted activity that cannot be captured in one overall indicator, not even in a composed-indicator whereby multiple indicators are weighted and summed. Evaluation of research activity requires a broad set of well-chosen indicators, each with its own meaning. It is also crucial that these indicators are well understood and – if necessary - explained to the possible users of the evaluation exercise. Additionally, each indicator can also be calculated for a speci c eld or sub eld in order to get a more in-depth view of the research activities of an institute.
• Similar Institutes: The comparability of research performance of institutes can be problematic in many instances. We claim that performance scores of an institute should be compared with scores of similar institutes where we de ne “similar” as being active in the same areas of science. A clustering of institutes on the basis of the research pro les was used to create eight different groups (Thijs and Glänzel, 2008). Each of these groups characterises institutes with a different research focus and different research behaviour. The groups are labelled by the elds in which they are most active. This label does not exclude (some lower) activity in other elds. These are the eight groups
1. Biology 2. Agriculture 3. Multidisciplinary 4. Geo & Space Science
5. Technical & Natural Sciences 6. Chemistry
7. General & Research Medicine 8. Specialised Medicine
We could show that the classi cation of institutes on the basis of their pro les creates 8 groups that differ signi cantly from each other on publication and citation indicators. (Thijs and Glänzel, 2009) It is a valid strategy to compare the performance of an institute with the average of the group. We can also divide all groups in quintiles in order to describe the position of an institute within the cluster (rank score 1 to 5). The combination of this principle with the evolution allows us to compare the improvements of an institute with the changes in other similar institutes.
4.3 Data sources and data processing
The data were extracted from the annual volumes from the Web of Science of Thomson Reuters (Philadelphia, PA, USA). For this study all original research documents (i.e. article, letter, note and review) with at least one Israeli address and indexed between 1992 and 2007 were examined.
The entire publication period of sixteen years was subdivided into six year ranges (1991-1993, 1994-1996, 1997-1999, 2000-2002, 2003-2005 and 2006- 2007) and citations received by papers were counted for a three-year citation- window each beginning with the year of publication. For the last set (2006-2007) the citation window is incomplete; therefore citation indicators are not calculated for this set.
Subject-classi cation of the publications was based on the eld assignment of journals according to sixteen major elds of science developed in Leuven and Budapest (Glänzel & Schubert, 2003). These elds are Agriculture & Environment, Biology (Organismic & Supraorganismic Level), Biosciences (General, Cellular &
Subcellular Biology, Genetics), Biomedical Research, Clinical and Experimental Medicine I (General & Internal Medicine), Clinical and Experimental Medicine II (Non-Internal Medicine Specialties), Neuroscience & Behaviour, Chemistry, Physics, Geosciences & Space Sciences, Engineering, Mathematics, Social Sciences I (General, Regional & Community Issues), Social Sciences II (Economical & Political Issues) and Arts & Humanities. Multidisciplinary journals form a separate group.
Publications were assigned to countries and Israeli institutes based on all the addresses that appear in the by-line of the paper using a three-step assignment procedure. First, a list of distinct names of institutes as occurring in the extracted addresses was compiled. This list contained all possible synonyms and spelling variance/errors of research institutes. Secondly, each entry in this list with a number of publications above a certain threshold was assigned, if possible, to a unique, known institute. Finally, the addresses in the by-line of the papers were matched to the entries in the thesaurus. This procedure resulted in the identi cation of 303 unique institutions.
After the data cleansing for each institute, a research pro le was calculated for each of the used publication windows. These pro les were then used to classify the institutes into one of the eight different groups.
4.4 Methods and application
In this study we apply the methodology of a dynamic perspective on research performance on ve Israeli academic institutions. We classify the most productive Israeli research institutions and compare their performance to national standards and benchmarks for each speci c group. We also examine the changes in the Israeli research landscape by studying the dynamics of the research institutes and relating them to changes in research systems. Each of the institutes is classi ed into one of the eight groups as mentioned above. In the two studies on classi cation of institutes (Thijs and Glänzel 2008&2009) we used 1767 institutes from EU15 without Greece but including Switzerland. This allows us to compare the performance of the Israeli institutes with likewise European institutes.
Several commonly used bibliometric indicators are used to gain insight into the differences between the institutes:
• Share of publications in Israel and the growth rate.
• Share of author self-citations (SCit%) as the percentage of author self- citations in all citations received in the same period.
• Mean Observed Citation Rate (MOCR) is de ned as the ratio of citation count to publication count. MOCR is here based on 3-year citation windows.
• Mean Expected Citation Rate (MECR) is a journal-based indicator expressing an expected citation rate of a given paper set. The (journal- based) expected citation rate of a single paper is de ned 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.
• Relative Citation Rate (RCR) is the ratio of MOCR and MECR. RCR thus compares the observed citation rate with the expected one. RCR
= 0 corresponds to uncitedness, 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.
• Relative Citation Rate eXcluding self-citations (RCRX) is the ratio of observed citation rate and expected citation rate, both with self- citations excluded. The interpretation of this indicator is analogous to the previous indicator.
• Normalized Mean Citation Rate. (NMCR) compares analogously to the RCR the observed citation rate (MOCR) with an expected citation rate.
Here, however, a eld based expected citation rate is used expressing the average number of citations received over all papers published within a speci c eld within one year. NMCR = 0 corresponds to uncitedness, NMCR < 1 means lower-than-average, NMCR > 1 higher-than-average citation rate, NMCR = 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 eld to which the papers belong.
• MECR/FECR gives the ratio between the two expected citation rates for a given set of publications. A value above 1 means that the journal based expected rate is higher than the rate based on the eld. As journals belong –at least partially- to the given eld this value higher than 1 corresponds to the fact that the set of papers is published in journals more visible than the eld in average. A value lower than 1 means thus that the papers are published in journals belonging to the lower segment within the eld.
• Share of international co-authored papers (%Int Coll) is the percentage of papers with more than one country in the by-line of the paper in all publications.
• Relative Citation Rate of International Co-Authored papers (RCR Int.
Coll) compares the observed citation rate of international co-authored papers to the expected citation rate of the international co-authored papers in the journal.
4.5 The macro level: Israel as a country
First, it is important to look at the national situation in Israel before we examine the individual institutes. In order to get a realistic picture of the growth of a national publication output one has to compare this with the overall growth of the underlying database. In particular, there are two methods for such comparison,
rstly by monitoring the evolution of the national share in the world total and, secondly, by directly comparing the evolution of the national growth with that of the world or other reference countries. We apply both methods. Figure 7 presents the growth rate both for Israel as for the world. This shows that the number of Israeli papers in the database grows in line with the overall growth in the world.
The rst column of table 27 (header ‘% Pub’) shows a constant share of Israeli papers in the database.
Fig. 7. Growth rate both for Israel and the world
Other indicators are also given in table 27.
Time % Pub MOCR MECR %Int Coll RCR Int Coll
1991-1993 1.1% 3.44 3.83 33.4% 1.09
1994-1996 1.2% 3.93 4.11 35.1% 1.20
1997-1999 1.2% 4.26 4.39 38.0% 1.15
2000-2002 1.2% 4.91 4.83 40.7% 1.21
2003-2005 1.2% 5.30 5.30 41.2% 1.17
2006-2007 1.1% 42.2%
Time %SC RCR RCRX NMCR MECR/FECR
1991-1993 31.0% 0.90 0.86 1.00 1.11
1994-1996 30.4% 0.96 0.93 1.09 1.14
1997-1999 28.6% 0.97 0.95 1.11 1.14
2000-2002 26.8% 1.02 0.99 1.17 1.15
2003-2005 25.1% 1.00 0.98 1.13 1.13
2006-2007
Table 27. Publication and Citation indicators for Israel
4.6 The meso level: Israeli institutes
We selected the ve most productive Israeli research institutes responsible for more than 70% of all publications in the respective time frames. Each of these institutes produces more than 10% of all Israeli papers in the last period.
• Tel Aviv University
• Hebrew University of Jerusalem
• Technion – Israel Institute of Technology
• Ben Gurion University of the Negev
• Weizmann Institute of Science
Tel-Aviv University
Hebrew University
Technion- Institute of technology
Ben Gurion University
Weizmann Institute of Science Agriculture &
Environment
1.3 6.4 3.5 5.2 1.7
Biology 6.0 12.9 3.5 9.7 6.1
Biosciences 12.0 13.2 8.2 8.0 22.9
Biomedical Research 5.6 5.7 4.3 4.0 3.0
General & Internal Medicine
15.6 6.2 9.2 9.0 9.4
Non-internal Medicine Specialties
24.3 8.5 12.6 12.4 3.0
Neuroscience &
Behavior
6.6 7.0 4.0 6.8 4.9
Chemistry 5.9 16.0 14.4 17.4 14.7
Physics 16.6 13.7 25.1 20.2 29.9
"Geosciences & Space Sciences"
4.7 5.9 4.7 4.4 3.8
Engineering 11.7 6.7 22.9 15.5 10.6
Mathematics 7.8 7.6 10.1 7.5 5.3
Multidisciplinary Sciences
1.3 1.8 1.0 0.6 4.6
Social Sciences I (General, regional and community issues)
2.9 5.4 0.8 2.7 0.7
Social Sciences II (Economical & Political issues)
3.7 5.1 1.5 4.5 1.1
Arts & Humanities 1.9 3.0 0.1 1.7 0.0
Table 28. Research pro le of the 5 selected institutes (2003-2005) (in percentages)
In table 28 the research pro les for the last complete period are given (2003-2005). These pro les are share of publications for each eld in the total output of the institute. On the basis of these pro les the selected institutes could be classi ed as multidisciplinary institutes in each of the different time periods. The indicators for the Multidisciplinary Group in total can be found in the appendix.
4.6.1 Tel Aviv University
Tel-Aviv University was established in 1956. Today it is the largest university in Israel. It is comprised of nine faculties (arts, engineering, exact sciences, humanities, law, life sciences, medicine, social sciences and a business school ), 106 departments and 90 research institutes (http://www.tau.ac.il/intro- eng.html). In the 2005-6 academic year there were more than 28,000 enrolled students in Tel-Aviv University at all levels, about 10,000 masters students and 2,000 doctoral students . The university had 2,263 research staff members during the 2004-5 academic year .
Publications MOCR RCR
Share Position Value Position Value Position
1991-1993 23.8% 1 3.17 3 0.91 4
1994-1996 24.2% 1 3.78 3 0.98 4
1997-1999 22.8% 1 4.13 3 0.97 4
2000-2002 21.2% 1 4.5 3 0.97 4
2003-2005 22.4% 1 5.18 3 0.98 4
% Self Citations RCRX NMCR
Value Position Value Position Value Position
1991-1993 33.2% 3 0.85 4 0.92 3
1994-1996 32.8% 3 0.92 4 1.03 3
1997-1999 30.7% 3 0.92 4 1.05 3
2000-2002 29.2% 3 0.92 4 1.07 3
2003-2005 25.2% 2 0.96 4 1.08 3
Table 29. Indicators for Tel Aviv University
Table 29 gives values and positions for six indicators for Tel Aviv University. The column with the ‘Value’-header gives the calculated value, in the next column (heading ‘Position’) the quintile position of the institute is given within the Multidisciplinary cluster for this indicator (550 European multidisciplinary institutes and all Israeli multidisciplinary institutes are taken into account to calculate the position). For the publication count, the value of total number of publications is not given, only the quintile position. Instead, the share of the institutes’ publications in the total output of Israel is given.
Tel Aviv University is without doubt the most proli c research institute in Israel producing more than one paper in ve. The share remains quite stable over time. In the multidisciplinary group it is positioned in the rst quintile with respect to paper output over the ve different time periods. But so are the four other institutes considered in this study.
In contrast to the position on publication output, Tel Aviv University never appears in the top quintile. The Mean Observed Citation Rate is increasing over time but is in the third quintile. The Relative Citation Rate is always below 1 resulting in the placement of Tel-Aviv University in quintile 4 for this indicator.
This means that Tel Aviv University attracts on average less citations than should be expected on the basis of the journals in which its articles are published.
The Normalized Mean Citation Rate is nicely improving over time with only in 1991-1993 a value below 1. This means that Tel Aviv University has a better citation rate than the eld expected citation rate. Combining this and the previous observation we can conclude that Tel Aviv University selects on average journals with a higher expected citation rate than the eld average. This is also con rmed by the results shown in Figure 8 and gure 9. These relational charts (Glänzel, 2009) are designed for the presentation of normalized indicators. The chart area is divided in 6 sections by adding three lines: x=1, y=1 and x=y. A value on the vertical axis above 1 corresponds with an observed citation rate higher than the expected based on the eld assignments. A horizontal value at the right side of 1 means that on average the expected citation rate of the journals in which the institute publishes is higher than the average expected citation rate of the eld. This means that high impact journals are selected. An x-value higher than the y-value corresponds to a RCR higher than 1, thus an observed citation rate higher than the journal expected citation rate. Figure 8 and Figure 9 each have a 5 year publication window (1991-1995; 2001-2005). In the rst period Tel Aviv University is located below 1 on the vertical axis and just right of the 1 on the horizontal. In the second period we see a shift to the right and a bit moving up. However Tel Aviv University is not above the x=y line. We can conclude that Tel Aviv university can publish in high impact journals and in the second period has a higher citation than the eld average.
4.6.2 Hebrew University
The Hebrew University of Jerusalem was established in 1925. The university has seven faculties (humanities, law, social sciences, science, medicine, dental medicine and agriculture) and several schools (http://www.huji.ac.il/huji/
eng/aboutHU_e.htm). Its research staff had 2,031 members in the 2004-5 academic year, and the number of enrolled students in 2005-6 was almost 22,000 including about 7,000 masters and 2,700 doctoral students.
This institute’s share is declining over time from nearly one in ve papers to one out of seven publications in Israel. As mentioned above, the position of all
ve institutes on publications is in the rst quintile. The share of internationally co- authored papers is also increasing from 38% to 48%, and the Hebrew University is placed either in the rst or the second quintile for this indicator, depending on the period under consideration. Looking at the citation indicator we see a good MOCR with a value increasing from 4.19 to 6.07 positioning in the 2nd group.
The MECR/FECR ratio higher than 1.2 and in the rst group indicates a clear
distinction between the expected citation rate of the journals versus the expectancy based on the journal. The RCR values in quintiles 4 or 3 and the NMCR in quintile 2 lead us to believe that the Hebrew University has a strategy to aim at journals with high visibility in the eld. However, their papers are not able to attract as many citations as those based on the journal average.
Publications MOCR RCR
Share Position Value Position Value Position
1991-1993 19.3% 1 4.19 2 0.90 4
1994-1996 19.1% 1 4.76 2 0.95 4
1997-1999 15.9% 1 5.07 2 1.00 3
2000-2002 14.9% 1 5.60 2 1.03 3
2003-2005 14.4% 1 6.07 2 0.98 4
% Self Citations RCRX NMCR
Value Position Value Position Value Position
1991-1993 31.0% 2 0.86 4 1.09 2
1994-1996 31.7% 3 0.90 4 1.18 2
1997-1999 31.4% 3 0.95 3 1.26 2
2000-2002 29.9% 3 0.99 3 1.31 2
2003-2005 27.9% 3 0.93 4 1.26 2
Table 30. Indicators of Hebrew University
4.6.3 Technion – Israel Institute of Technology
The Technion, Israel Institute of Technology, was founded in 1924, and it is the oldest institute of higher education in Israel (http://pard.technion.ac.il/
fastfacts/FramsFactsE.asp?myret=main). Although it has a faculty of humanities, the emphasis is on engineering, science and medicine. It has eighteen faculties - most faculties at the Technion correspond to departments at the other institutions (aerospace engineering, architecture and town planning, biology, biomedical engineering, biotechnology and food engineering, chemical engineering, chemistry, civil and environmental engineering, computer science, education in technology and science, electrical engineering, humanities and arts, industrial engineering and management, materials engineering, mathematics, mechanical engineering, medicine and physics). The number of research staff members was 1,298 in 2004- 5, and the number of enrolled students in 2005-6 was about 12,500, with masters 2,900 and 880 doctoral students.
Most of the indicators for Technion show no real evolution. Only MOCR and share of self-citations demonstrate a clear increase/decline. The position in the
rst group of the MECR-FECR ratio is striking. Just as with Tel Aviv University and the Hebrew University, the Technion tends to choose journals with high visibility without being able to match the expected citation rates.
Publications MOCR RCR Share Position Value Position Value Position
1991-1993 14.5% 1 2.86 4 0.92 4
1994-1996 15.6% 1 3.26 4 0.97 4
1997-1999 15.1% 1 3.80 3 1.02 3
2000-2002 14.8% 1 4.02 3 0.94 4
2003-2005 15.5% 1 4.61 3 0.97 4
% Self Citations RCRX NMCR
Value Position Value Position Value Position
1991-1993 35.3% 3 0.87 4 1.03 3
1994-1996 33.7% 3 0.94 3 1.13 2
1997-1999 29.7% 2 1.02 3 1.23 2
2000-2002 30.4% 3 0.91 4 1.16 3
2003-2005 28.6% 3 0.93 4 1.14 3
Table 31. Indicators for Technion - Israel Institute of Technology
4.6.4 Ben-Gurion University
Ben-Gurion University of the Negev was founded only in 1962 (http://
cmsprod.bgu.ac.il/Eng/home/About/history.htm). It has four faculties (health sciences, humanities and social sciences, natural sciences), two schools (business administration and advanced studies) and several research institutes. In 2004-5, there were about 18,000 enrolled student at Ben-Gurion University, with 4,000 masters students and slightly less than 1,000 doctoral students.
Publications MOCR RCR
Share Position Value Position Value Position
1991-1993 8.3% 2 2.17 5 0.74 5
1994-1996 9.4% 2 2.53 4 0.85 5
1997-1999 9.7% 1 2.64 4 0.83 5
2000-2002 10.2% 1 2.94 4 0.86 5
2003-2005 10.5% 1 3.52 4 0.84 5
% Self Citations RCRX NMCR
Value Position Value Position Value Position
1991-1993 39.8% 4 0.64 5 0.74 4
1994-1996 37.8% 4 0.79 4 0.83 4
1997-1999 36.8% 4 0.77 5 0.83 4
2000-2002 34.8% 4 0.80 5 0.86 4
2003-2005 31.1% 4 0.79 5 0.87 4
Table 32. Indicators for Ben-Gurion University
Ben-Gurion University of the Negev is the smallest of the 5 institutes examined in terms of publication share. However, it is the only one with a clearly increasing share of publications. In the periods 1991-1993 and 1994-1996 it was positioned in the second group with respect to publication counts. The share of internationally co-authored papers is roughly average. Although the RCR for these international papers is higher than the overall RCR, Ben-Gurion is the only institute where the international papers are less visible than what might be expected, based on the journals. The MECR-FECR ratio is around 1 so both expected citation rates are in line with each other. The share of author self-citations is above the average and the RCR with self-citations excluded drops far below 1.
4.6.5 Weizmann Institute of Science
The Weizmann Institute of Science, unlike the other institutions studied in this paper, is not a university but a multidisciplinary research institution. There are only graduate studies at Weizmann, in 2005-6 it had 280 masters and 680 doctoral students and 400 research staff members. The Weizmann Institute of Science was founded 1934 (http://www.weizmann.ac.il/acadaff/Scienti c_Activities/current/
weizmann.html).
The results for the Weizmann Institute of Science are impressive, especially when considering its size. For ve indicators the Weizmann Institute is positioned in the top group for all studied time-windows. Only 2 values are positioned in group 3 (RCR in 1991-1993, RCR Int Coll in 1997-1999). RCR values are well above 1. The share of self-citations is quite low and the RCR excluding self- citations is similar to the overall RCR. The MECR-FECR ratio with values near 1.5 indicate that the Weizmann Institute aims at the most visible journals, but their papers perform even better than the expected citation rate for these top journals.
Publications MOCR RCR
Share Position Value Position Value Position
1991-1993 12.0% 1 7.86 1 1.08 3
1994-1996 11.5% 1 8.89 1 1.14 2
1997-1999 11.1% 1 8.67 1 1.10 2
2000-2002 11.2% 1 10.11 1 1.15 2
2003-2005 10.4% 1 10.38 1 1.17 2
% Self Citations RCRX NMCR
Value Position Value Position Value Position
1991-1993 26.6% 1 1.06 2 1.51 1
1994-1996 25.9% 1 1.13 2 1.74 1
1997-1999 25.4% 1 1.09 2 1.67 1
2000-2002 24.3% 2 1.13 2 1.79 1
2003-2005 22.9% 2 1.17 2 1.75 1
Table 33. Indicators for Weizmann Institute
Fig. 8. MECR/FECR vs NMCR 1991-1995
Fig. 9. MECR/FECR vs NMCR 2001-2005
4.7 Conclusions
The meso analysis of Israel’s big universities and research institutes brought new insight into the national research landscape. Beyond doubt, most institutes follow the national trends; international co-authorship is considerable in
all institutes under study and the share of author self-citations decreases steadily.
On the other hand, the normalised mean citation rate increases indicating the growing impact of Israeli research. Weizmann Institutes holds the most favourable position, however, also Technion and Hebrew University hold positions distinctly above the world standard over the entire period. And the same holds for Tel Aviv University in the new millennium as well. In all, bibliometric indicators re ect a positive development for the ve largest Israeli research institutions. The breakdown by subject areas and the in-depth analysis of faculties, departments and research groups could bring further insight into Israel’s science system, but this is left for future research.
References
ARWU (2007), Academic Ranking of World Universities, accessible via: http://
ed.sjtu.edu.cn/ranking.htm.
Butler, L. (2007), Assessing University Research: A Plea for a Balanced Approach, Science and Public Policy, 34(8), 565-574.
CHE (2007), CHE University Ranking, accessible via: http://www.che-ranking.
de.
Glänzel, W., Schubert, A. (2003), A new classi cation scheme of science
elds and sub elds designed for scientometric evaluation purposes, Scientometrics, 56 (3), 357–367.
Glänzel, W., Debackere, K. (2008), On the ‘multidimensionality’ of ranking and the role of bibliometrics in university ranking, ULB, Brussels, to be published.
Leta, J., Glänzel, W., Thijs B., Science in Brazil. Part 2: Sectoral and institutional research pro les. Scientometrics, 2006, 67 (1), 87-105.
THES-QS (2007), Times Higher Education - QS World University Rankings 2007, accessible via: http://www.topuniversities.com/
worlduniversityrankings/results/2007/overall_rankings/top_400_
universities/.
Thijs, B., Glänzel W. (2008), A structural analysis of publication pro les for the classi cation of European research institutes. Scientometrics, 74 (2), 223-236.
Thijs, B., Glänzel W. (2009), A structural analysis of benchmarks on different bibliometrical indicators for European research institutes based on their research pro le. Scientometrics, 79(2), 377-388.
van Raan, A. (1999) Advanced bibliometric methods for the evaluation of universities. Scientometrics, 45 (3),417-423.
Zimmerman, E., Glänzel, W., Bar-Ilan, J. (2008), Scholarly collaboration between Europe and Israel: A scientometric examination of a changing landscape, Scientometrics, 78(3), 427-446.
Appendix I
Mean values of indicators by pro le clusters
Indicator BIO AGR MDS GSS TNS CHE GRM SPM
Publications 248.5 113.8 1226.4 150.8 528.1 87.3 120.3 209.4
Citations 1159.7 387.0 7065.2 1026.8 2140.9 305.0 927.8 1200.2
Cited% 78% 75% 77% 78% 66% 65% 78% 71%
SCit% 32% 33% 28% 37% 37% 32% 19% 19%
Coop% 44% 34% 39% 65% 44% 36% 29% 23%
MOCR 4.41 3.21 5.41 5.41 3.27 3.07 6.55 4.68
MOCRX 3.12 2.18 4.03 3.41 2.15 2.16 5.30 3.78
MECR 4.15 2.94 4.82 5.06 2.94 3.08 5.62 4.05
MECRX 3.03 2.03 3.63 3.53 1.94 2.02 4.60 3.28
RCR 1.04 1.09 1.07 1.04 1.07 0.97 1.14 1.13
RCRX 1.00 1.07 1.05 0.95 1.05 1.04 1.13 1.13
NMCR 1.08 1.04 1.12 1.16 1.11 0.90 1.10 1.01
NMCR / RCR 1.03 0.96 1.02 1.10 1.02 0.93 0.95 0.88
Source: Thijs and Glänzel (2009)
