Bayesian correlated factor analysis of socio-demographic indicators |
| |
Authors: | Maura Mezzetti Francesco C. Billari |
| |
Affiliation: | (1) Institute of Quantitative Methods and IGIER, Universitá Bocconi, Milano, Italy |
| |
Abstract: | Recent changes in European family dynamics are often linked to common latent trends of economic and ideational change. Using Bayesian factor analysis, we extract three latent variables from eight socio-demographic indicators related to family formation, dissolution, and gender system and collected on 19 European countries within four periods (1970, 1980, 1990, 1998). The flexibility of the Bayesian approach allows us to introduce an innovative temporal factor model, adding the temporal dimension to the traditional factorial analysis. The underlying structure of the Bayesian factor model proposed reflects our idea of an autoregressive pattern in the latent variables relative to adjacent time periods. The results we obtain are consistent with current interpretations in European demographic trends. |
| |
Keywords: | Bayesian inference Factor analysis Correlated factor loadings European family dynamics Demographic convergence |
本文献已被 SpringerLink 等数据库收录! |
|