Pitfalls in Aggregating Performance Measures in Higher Education

19 Pages Posted: 3 Mar 2017

See all articles by Ross Williams

Ross Williams

University of Melbourne - Melbourne Institute: Applied Economic & Social Research

Gaétan de Rassenfosse

École Polytechnique Fédérale de Lausanne

Date Written: January 1, 2016

Abstract

National and international rankings of universities are now an accepted part of the higher education landscape. Rankings aggregate different performance measures into a single scale and therefore depend on the methods and weights used to aggregate. The most common method is to scale each variable relative to the highest performing entity prior to aggregating. Other approaches involve transforming the data to allow for the different spread of the variables. We evaluate alternative methods and the sensitivity to weights with illustrations from the Times Higher Education and Shanghai Jiao Tong rankings of universities and the U21 rankings of national systems of higher education. We conclude that transforming the data clouds interpretation; the choice of included variables is more important than the weights attached to them; and there are limitations in extending ranking to a large number of universities/countries.

Keywords: Higher Education System; Ranking; Universitas 21; Performance Measures; Shanghai Ranking; Times Ranking

Suggested Citation

Williams, Ross and de Rassenfosse, Gaétan, Pitfalls in Aggregating Performance Measures in Higher Education (January 1, 2016). Available at SSRN: https://ssrn.com/abstract=2926246 or http://dx.doi.org/10.2139/ssrn.2926246

Ross Williams

University of Melbourne - Melbourne Institute: Applied Economic & Social Research ( email )

Level 5, FBE Building, 111 Barry Street
Parkville, Victoria 3010
Australia

Gaétan De Rassenfosse (Contact Author)

École Polytechnique Fédérale de Lausanne ( email )

Station 5
Odyssea 1.04
1015 Lausanne, CH-1015
Switzerland

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
27
Abstract Views
368
PlumX Metrics