Goal-based participatory weighting scheme: balancing objectivity and subjectivity in the construction of composite indicators

• Published in: Quality & quantity Vol. 57; no. 5; pp. 4387 - 4407
• Main Authors: Correa Machado, Alexei Manso, Ekel, Petr Iakovlevitch, Libório, Matheus Pereira
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.10.2023; Springer; Springer Nature B.V
• Subjects: Methodology of the Social Sciences; Objectives; Objectivity; Researcher subject relations; Social Sciences; Subjectivity; Validity; Weighting; Composite indicators; Multi-criteria methods; Social exclusion; Optimization; Multidimensional phenomena
• ISSN: 0033-5177; 1573-7845
• DOI: 10.1007/s11135-022-01546-y

The weighting of sub-indicators is a controversial topic in the literature, being the object of investigation by many researchers. The absence of a clear definition for the most appropriate sub-indicator weighting scheme has led researchers to combine subjective and objective weighting schemes. In this work, an innovative sub-indicator weighting scheme, the Goal-based Participatory weighting scheme, is proposed. It employs the Generalized Reduced Gradient (GRG) to achieve two. Objectives: First, to construct composite indicators more correlated with the variable of greater conceptual significance in the multidimensional phenomenon, ensuring its external validity; Second, to define weights for the sub-indicators respecting the opinion of the group of experts, ensuring its consensus degree. Eleven composite indicators for the Equal Weights (EW), Participatory and Data-driven weighting schemes were constructed by the Technique for Order Preferences by Similarity to Ideal Solutions (TOPSIS). The results show that the Goal-based Participatory weighting allows for a more balanced solution regarding external validity and consensus degree when compared to other weighting schemes. The external validity obtained by the Goal-based Participatory weighting is greater than the one obtained by the EW, Participatory, and Data-driven weighting in 91% of the composite indicators analyzed, being 30% higher in 15% of the composite indicators and 37% higher than one of the composite indicators constructed by the Participatory weighting. All composite indicators constructed using the GRG showed a higher consensus degree when compared to composite indicators constructed by the EW and Data-driven weighting schemes.