A robust composite indicator framework for evaluating Sustainable Development Goal 3 using benefit-of-the-doubt models and principal component analysis

Abstract

This study proposes a novel benefit-of-the-doubt (BOD) model for constructing composite indicators (CIs) to assess Sustainable Development Goal 3 across the World Health Organization (WHO) member states. To address the BOD model's limited ability to differentiate among states with many sub-indicators, we introduce a common weight BOD (CWBOD) model to improve cross-country comparability. To improve the model's discriminatory power, we apply principal component analysis (PCA) to reduce the number of sub-indicators, using the resulting principal components as inputs to the BOD model. To account for the uncertainty due to possible information loss in PCA, we further develop a robust BOD (RBOD) model. The final CI scores are computed using the geometric mean of the BOD, CWBOD, and RBOD scores. We apply this integrated framework to compute a Public Health Index for 177 WHO member states, enabling a more precise and robust evaluation of global public health performance.