A0756
Title: Quantitative robustness assessment of composite indicators: Methodology and application
Authors: Viet Duong Nguyen - University Carlo Cattaneo - LIUC (Italy) [presenting]
Abstract: Benchmarking exercises based on composite indicators are widely used in decision-making across diverse domains, ranging from policy formulation to performance evaluation. However, producing a reliable benchmark poses significant challenges in multi-criteria decision analysis, where outcomes may vary substantially due to minor changes in methods or input features. This uncertainty impacts both developers who strive to generate trustworthy comparisons and decision-makers who rely on these results for strategic planning. As a remedy, a novel approach is introduced for quantifying the robustness of composite indices, accounting for multiple sources of uncertainty such as normalization techniques, weighting schemes, and aggregation functions. The method offers a comprehensive evaluation of index performance based on a model configuration grid, designed to identify the interacting factors that significantly compromise pairwise ranking stability. Its utility is demonstrated through a measuring application using the environmental sustainability dashboard (UNDP), where the proposed method provides valuable insights into benchmarking consistency across computational schemes and highlights critical points for robustness improvement.
