ABSTRACT
Variance decomposition using the analysis of variance is an estimable function and could have been placed in the previous chapter that focuses on estimable functions. It appears here, however, because it is closely related to the age-period-cohort mixed model (APCMM) method of determining the variance associated with random factors that is not an estimable function. Both the analysis of variance approach and the APCMM approach are excellent ways to wrest information about the unique effects of ages, periods, and cohorts from APC data. The APCMM is also related to the hierarchical age-period-cohort (HAPC) model of Yang and Land (2006). I will mention several of these relationships in this chapter that should help readers gauge some of the advantages and limits of these approaches.
