Modelling the latent components of personal happiness

We discuss a class of statistical models able to measure the self-evaluation of happiness by means of a sample of respondents and investigate the ability of this proposal to enhance the different contribution of subjective, environmental and economic variables. The approach is based on a mixture model introduced for interpreting the ordered level of happiness as a combination of a real belief and a surrounding uncertainty: these unobserved components may be easily parameterized and immediately related to subjects' covariates. An empirical evidence is supported on data set derived by the Survey of Household Income and Wealth (SHIW) conducted by the Bank of Italy.