Subjective Well-Being over the Life Span: Modeling Age-, Period-, and Cohort-Effects in the Additive Mixed Model Framework

The analysis of panel data is often concerned with effects of time-related changes. Dealing with data on individuals, the variation of some dependent variable one is interested in can be associated to differences in the specific ages of the individuals, the different birth-cohorts they belong to, and changes in the period of observation. Because of the the Age-Period-Cohort Identification prob-lem it is, however, in general not possible to estimate distinct effects for each of the three covariates. This thesis provides an approach for analyzing such data in the additive mixed model framework. Estimating two smooth effects, the age-effect and the cohort-or period-effect, and additionally an interaction surface of both it can be argued that the model takes all three time-related changes into account. This modeling approach is illustrated by an application on german socio-economic data regarding subjective well-being over the life span and various graphics for an evaluation of such models are presented in this context