Tuesday, 29 April 2014

Choosing aggregation methods - A comparative analysis

Choosing aggregation methods - A comparative analysis
Liv Bjerre, Friederike Römer, Malisa Zobel (WZB)

This presentation sets out to compare three aggregation methods: a standard additive aggregation function, a geometric aggregation function and a non-compensatory/non-linear aggregation function brought forward by Munda & Nardo (2006). In its theoretical part this paper discusses the underlying assumptions and requirements for each of the aggregation methods. In the empirical part the three methods are applied to the IMPIC Index, a new dataset which measures immigration policies restrictiveness across all 33 OECD countries for the period 1980-2010. Furthermore, it is analyzed how the use of weights affects the score/rank order of the cases. The aim of the paper is thus twofold: it shall help researchers to choose an aggregation function adequate for their research question while at the same time adding to the overall discussion of robustness and sensitivity of indices which is taking place across disciplines.

Resources:

Munda, Giuseppe, and Michela Nardo (2009) 'Noncompensatory/nonlinear composite indicators for ranking countries: a defensible setting', Applied Economics 41(12): 1513-1523.