Webometrics Ranking of World Universities

1
Webometrics Ranking ofWorld Universities
3rd Meeting of the International Rankings Expert
Group (IREG-3) Shanghai Jiao Tong University,
October 28 - 31, 2007
Isidro F. Aguillo Cybermetrics Lab (CSIC),
Madrid, Spain
2
Agenda

• New role of the Web in the scholarly
  communication
• Rankings advantages and disadvantages
• Collecting Web data and building indicators
• Coverage Not only World-Class Universities
• Search engines, the new intermediaries
• Institutional web domain uses and abuses
• WR, the combined indicator
• The Ranking Current results and future
  developments
• Structure of the ranking
• Problems Gaps, duplicates, extra indicators
• Results Correlation
• Success? Expected and actual visitors
• Academic digital divide

3
About Us

• CSIC, Spanish National Research Council
• The largest public research body in Spain
• 126 institutes devoted mainly to basic science
• 20 of the total Spanish scientific output
• 2 Nobel Prizes
• Cybermetrics Lab
• Research group at Center of Scientific
  Information and Documentation
• Editors of e-journal Cybermetrics
• EU projects EICSTES and WISER
• MAVIR network
• Consulting services to universities
• Staff Coordinator, 4 PhD, 1 technicians

4
About the Ranking

• Primary objective was not to rank institutions
• Promote web publication
• Support Open Access initiatives
• Current design
• Larger coverage than other similar Rankings
• Including developing countries institutions
• Not focusing only in research results
• Activity, visibility, impact, prestige, and
  quality better reflected in the Web presence
• Paying attention to Web bad practices
• Web Publish or Perish

5
Importance of the Web Presence

• The new Academic global market
• Increased mobility of professors and alumni
• International competition for human resources and
  funding
• Getting prestige and visibility in the digital
  world
• Web is the best showcase for Universities
• All missions covered teachingresearchtransfer
• Intangibles and the web contents
• Freedom of teaching
• Self-organisation and maturity
• Access to resources
• Scholarly (formal and informal) communication
• Universal target groups
• Colleagues all over the world
• Prospective students worldwide
• Economic interested stakeholders
• General (huge) audiences
• Richer and diversified contents at cheaper costs

6
Metrics of the Web

• Contents size
• Correlated with the knowledge production
  capabilities of the institution
• number of potential authors (faculty members,
  staff, alumni)
• available (access to) resources
• internal and external policies
• Quality resources
• Difficult criteria
• authoritativeness of the scholars (university)
• peer review 2.0
• formats (file types), language (lingua franca)
• Visibility
• Hypertext links networks
• New motivations, many linkers (siters)
• Open versus Closed (international) impact
• Popularity
• Users, visits, behavior, evolution, referrers
• Methodological problems (global comparison no
  feasible)

7
Rankings Advantages and Shortcomings

• Advantages
• Combine many variables
• Even heterogeneous variables
• Easy to understand
• Temporal series
• Disadvantages
• Simplistic
• Weighting usually arbitrary
• Potential users undefined
• Rankings in the Web
• PageRank algorithm (ranking web pages by a
  weighted visibility indicator) is key for
  explaining the success of Google search engine
• TrafficRank Ranking of web domains according to
  the number of visits intercepted by the Alexa
  system
• Model for a combined indicator
• Webometrics Rank (derived from Web Impact Factor)
• VisibilitySize (11)

8
World Rankings

• Biases in the indicators
• Focus on research (bibliometric data)
• Limited info (lack of standardization) about
  inputs (personnel, funding)
• Reputation is very subjective, dependent of
  population surveyed
• University is a complex, multifaceted institution
• Web indicators reflects teaching, research and
  knowledge transfer missions
• But, they do not measure them separately

9
Webometrics Ranking

• First published in 2004
• Since 2006, two editions (January and July)
• World coverage
• Higher education Institutions (mostly
  universities) 13,000 (Jul07)
• RD related institutions (mostly research
  centers) 4,500 (Jul07)
• Web indicators
• Only institutions with an independent web domain
• univname.edu not hosting.edu/univname (except
  Helsinki University)
• Data from search engines (web visibility
  intermediaries)

Indicators used in the calculation of the
combined indicator WR
10
Methodology

• Normalization
• log (na1) N search engine (Google, Yahoo,
  Live, Exalead)
• Na ----------------------------- a web
  domain
• log (max(ni)1)
• Median method (Size)
• Sa ½ ((Ga Ya La Ea) max (Ga,Ya,La,Ea)
  - min (Ga,Ya,La,Ea))
• Rich files
• Ra PDFa DOCa PPTa Psa
• Ranking
• Sa -gt Ra(S)
• Ra -gt Ra(R) Va -gt Ra(V) 11 WR (4V)
  (2S1R1Sc)
• Sca -gt Ra(Sc)

11
(No Transcript)
12
Expected Results

• North America
• Stanford University (1st region world), MIT
  (2nd), University of California, Berkeley (3th)
  and Harvard University (4th)
• University of Toronto (28th world)
• Europe
• Cambridge (1st region), Oxford (2nd), Zurich
  (3rd)
• Asia/Pacific
• Tokyo (59th world), Australian National (60th),
  National Taiwan (96th), New South Wales (112th),
  Kyoto (116th) and Beijing (120th) universities
• Latin America
• Universidad Nacional Autónoma de México (68th
  world)
• Universidade de Sao Paulo (128th world)

13
Unexpected Results

• Up
• Pennsylvania State University (5th world)
• CiteSeer, index of computer and information
  science papers
• University of Wisconsin, Madison (8th world)
• Internet Scout Project
• Linkoping University (8th Europe, 62nd world)
  
• Lysator, the Academic Computer Society
• Universitat Trier (9th Europe, 64th world)
• DBLP, Computer Science Bibliography
• Down
• Cornell University (7th world)
• Arxiv (www.arxiv.org, not under cornell.edu)
• Yale University (35th world)
• Princeton University (38th world)
• California Institute of Technology (42th world)
• Jet Propulsion Laboratory (jpl.nasa.gov, not
  under caltech.edu)
• Johns Hopkins University (72th world)
• School of Medicine (hopkinsmedicine.org, not
  under jhu.edu)

14
Bad practices
15
(No Transcript)
16
Comparative Analysis Top 200
17
Comparative Analysis Top 500
18
Academic Digital divide
Number of universities by country in the Top 200
of the Webometrics Ranking (July 2007)
19
Regional distribution
Number of universities by region in the
Webometrics Ranking (July 2007)
20
Lessons
Webometrics is the only academic ranking that
provides indicators for universities worldwide
and not only the World- Class ones Information
provided could be useful to increase the
commitment to electronic publication, guiding the
Web policies and therefore improving Web presence
and visibility Web rank is probably highly
related to the global level of activities,
prestige and overall performance of the
university Web indicators are not able to
discriminate relative contributions but they are
a reliable and feasible way to describe the full
picture Results shown that there is an
unexpected academic digital divide as top
European and Japanese universities appear in
delayed positions when comparing with their North
American counterparts
21
Thank You!Questions?Isidro F. Aguillo, José
L. Ortega, Mario FernándezCybermetrics
LabCINDOC-CSICJoaquín Costa, 2228002
MadridEspañawww.webometrics.infoE-Mail
isidro_at_cindoc.csic.es