Validation of multidimensional performance assessment models using hierarchical clustering

This study aims to develop a validation method for multidimensional performance assessment models using hierarchical clustering. The primary focus is to explore the construct validity of these models, ensuring their effectiveness in diverse settings. The validation method consists of deconstructing the model into discrete units called “cubicles”, each representing the convergence of distinct performance assessment dimensions, implementing hierarchical clustering and conducting a comprehensive cluster analysis. This approach allows for an in-depth examination of the models’ internal structure to ensure the construct validity. The approach also emphasises an iterative process for refining the validation method based on the clustering outcomes. The validation method’s efficacy is demonstrated by applying it to a three-dimensional sustainability assessment model, which unveils cohesive clusters of model cubicles, organised by their sustainability dimensions. This application underscores the method’s applicability for construct validation, offering valuable insights into multidimensional performance assessment models. The research contributes to the field by providing a methodological validation framework that can be adapted for various performance models, regardless of their application field offering insights into their structural validity and potential areas for refinement.