ABSTRACT
This chapter illustrates the applications of the family of hierarchical age-period-cohort (HAPC) models for more advanced analysis of repeated cross-sectional data designs in the forms of both sample surveys and aggregate rates. In addition to a new example using repeated cross-sectional surveys from the General Social Survey on happiness, the chapter continues to analyze the cancer mortality rate data to relate their temporal trends to potential risk factor mechanisms. It describes the extensions to HAPC models useful for solving substantive problems in empirical research. These extensions involve more advanced statistical methods, such as Bayesian methods and heteroscedastic regressions. The HAPC analysis of happiness suggests that life course patterns, time trends, and birth cohort differences in happiness are distinct and independent of each other. In applying the Bayesian method to the estimation of HAPC-Cross-Classified Random Effects Models, one can take steps to assess how the choice of priors may affect posteriors.
