Model-Based Socio-Economic Health Measures Using Causal Modeling

• Activity Number: 475 - SPEED: Predictive Analytics with Social/Behavioral Science Applications: Spatial Modeling, Education Assessment, Population Behavior, and the Use of Multiple Data Sources
• Type: Contributed
• Date/Time: Wednesday, August 1, 2018 : 8:30 AM to 10:20 AM
• Sponsor: Social Statistics Section
• Author(s): F. Swen Kuh* and Anton H. Westveld and Grace S Chiu
• Companies: Australian National University and Australian National University and Australian National University
• Keywords: Bayesian; causal; LHFI; socio-economic health; latent; MCMC

Abstract

This research attempts to develop model-based socio-economic health measures using causal inference and modelling. Many conventional ways of constructing an alternative measure to Gross Domestic Product (GDP) for a country's socio-economic health involve combining different observable metrics to form an index. However, the 'health' of a society is inherently latent, with the metrics being observable indicators of health. None to our knowledge so far in existing efforts that provide this alternative measure uses a model-based (latent health factor index (LHFI)) approach to reflect the latent health. This framework integratively models the relationship between metrics, the unobservable latent health, and the covariates that drive the notion of health. Moreover, we are extending the LHFI approach by integrating it with causal modelling to investigate the causes and effects embedded in the factors. We implement our model using data pertaining to different aspects of societal health and potential explanatory variables. The approach is structured in a Bayesian hierarchical framework and the results obtained by applying Markov Chain Monte Carlo (MCMC) techniques.