A new participatory framework to build and interpret composite indicators - An application to country competitivenessFunded under: FP7-JRC
Statisticians often argue that all the work done during the composition of a composite indicator, to collect data and to edit this data, seems to be �hidden� or �wasted�. On the other hand stakeholders like the idea that a composite indicator is capable of summarizing complex and or difficult concepts into one single number. The research behind this thesis will start by changing the process of obtaining a composite indicator. At first a composite indicator will be obtained by eliciting stakeholders� preferences. After obtaining the composite indicator the work that statisticians put into collecting and editing data, is used to identify the drivers behind the concept of the composite indicators, using regression. The methods to elicit stakeholders� preferences are not new and are known as paired comparison analysis. However new models are used to obtain an index out of the paired comparison data. These new models make use of Probabilistic Inversion and use IPF or PARFUM to solve them. Also these new models make no assumptions about the distribution of the scores of the alternatives as the traditional methods do, but obtain better variances estimated and covariance.
Bibliographic Reference: EUR 22492 EN (2006), 75 pp. Free of charge
Availability: Katalogue Number: LB-NA-22492-EN-N The paper version can be ordered online and the PDF version downloaded at: http://bookshop.europa.eu
Record Number: 200719439 / Last updated on: 2007-09-18
Original language: en
Available languages: en