Recent Developments in THES QS World University Rankings

1
Recent Developments inTHES QS World University
Rankings

• Ben SowterHead of ResearchQS

IREG-3 Shanghai Jiao Tong University 29
October 2007
2
A year of joy

• Engagement activities led in
• HK, Korea, Japan, Australia, Netherlands,
  Thailand, Taiwan, Singapore, Malaysia, NZ,
  India, US, Italy, UK
• Much feedback received leading to some
  significant developments

3
Peer Review

• Reviewers prevented from selecting their own
  institution
• Impact on institutions from countries with good
  response but low number of institutions from
  which to select

4
Implementation of online system

• Enable institutions to interact with our data
• Submission
• Validation
• Track history
• Explanatory notes fields

5
Full-Time Equivalent (FTE)

• Consistent usage of FTE data for all faculty and
  student data
• Requesting faculty numbers is complex
• Response very diverse
• FTE data requested in 2006 but not consistently
  supplied
• Request for both Headcount and FTE this year
  provided clarity
• Time spent explaining FTE concept to stakeholders
• Missing FTE data extrapolated using national
  averages and Headcount data

6
Switch to Scopus
7
Switch to Scopus

• No viable alternative to Thomson in 2004
• ESI not a thorough interpretation of Web of
  Science
• Work with Evidence/Thomson led to many
  universities missing data
• Main benefits of WoS relate to older data
  particularly pre-1996
• Scopus stronger journal representation outside US
• Scopus tracks some non-English content
• Access to much greater data detail

8
The Scopus Challenge

• Pre-processed over 50Gb of data from Scopus
• Manually matched over 14,500 name variants of 566
  universities
• This years citation counts correlate with last
  years ESI data at 0.89

9
Z-Score Aggregation

• Method used in various domestic rankings such as
  The Times (UK)
• Stabilise data year on year
• More accurate application of weightings
  throughout the sample
• Smoothes any anomalies that remain
• Essentially smoothes the curve for each
  individual indicator

10
The Effect of Z-Score Application on Rankings
Indicators
11
Effects

• Online System
• More thorough and complete data all the way down
  the list
• Less need for averages
• Peer Review
• Countries with disproportionately high response
  but a small number of institutions lose unnatural
  advantage
• FTE Data
• Universities with particularly high or low
  numbers for part-time will be influenced
• Scopus
• Less pronounced bias towards US and English
  speaking world
• Z-Score aggregation
• Less advantage for the best institutions in lower
  weighted indicators
• Increased stability year on year

12
2007 Indicator Correlations
13
Increased Response
14
Results Published November 8

• THES publish on November 8th
• November 9th results available on
  topuniversities.com
• Top 200
• Top 100 for each of 5 subject areas
• Top 100 by region
• More detail on methodology
• Increasingly detailed profiles of institutions
• Links to other rankings

15
Thank You

• Ben Sowter
• ben_at_qsnetwork.com