Multiversal model of measurement of a composite index of well-being

The aggregation of many indicators into a unique index may hide the latent uncertainty in the rank orders of the statistical units due to the freedom in the choice of the aggregating methods. Multiversal modelling enhance this flexibility in the operative definition to derive a posterior distribution of estimates for the latent composite index. This study applies this principle to map the distribution of quality of life in 107 provinces in Italy.