The Novel Minimal Uncertainty with Subjective Constraints Method for Constructing Composite Indicators and its Application for Analyzing Child Well-Being and the Health System

Abstract This paper introduces the Minimum Uncertainty Method (MinUn-M), a new approach for constructing composite indicators that integrate subjective expert opinions with objective, data-driven techniques to analyze multidimensional phenomena that are not measurable by a single indicator. The MinUn-M method provides a stable composite indicator in which variations in the model inputs do not lead to significant changes in the model outputs, ensuring greater reliability in representing multidimensional phenomena and facilitating informed decision-making by managers and policymakers. The methods also minimize incompatibilities between subjective and objective weights, thereby promoting alignment between theory and practice by treating the expert’s opinions as constraints on the objective function that minimizes the uncertainty of the composite indicator ranks. Through two case studies, the MinUn-M method is applied to address the challenges of measuring and correlating the child well-being and health systems in low- and middle-income countries, as well as child well-being in Brazilian states. The results demonstrate superior robustness and precision compared to traditional weighting methods (e.g., equal weights, factor analysis, entropy). The main methodological contribution is the balance between objectivity and subjectivity in allocating weights, providing practical tools for public management and future research on constructing multidimensional indicators.