Hur kan vi bedöma tillförlitligheten i en rangordning?- en Bayesiansk ansats
English title: How can we evaluate the reliability of rankings?
Author(s): Leif Nordberg -
Type: Journal article
During recent years it has become quite popular to rank units like countries, universities or researchers according to a variety of indicators. For example, the so called Shanghai ranking of almost all universities in the whole world has received a lot of attention. On the other side, this ranking has also been the subject of fierce attacks. Bayesian methods provide one opportunity to study the uncertainty in rankings due to the fact that the observed values of the indicators on which the rankings are based are more or less wrong. The use of Bayesian methods is illustrated with two examples of which the first is related to the measurement of the support of political parties and is based on data for Finland from the 2010 European Social Survey. The second example concerns the measurement of publication activity in the social science faculties in the universities in Finland 2009. The results show that even if the top- and bottom-rankings are quite stable, the middle-rankings may be much more uncertain.
From page no: 94
To page no: 102
Journal: Ekonomiska samfundets tidskrift