A New Age-Period-Cohort Model for Describing and Investigating Inter- and Intra-Cohort Effects

Liying Luo, University of Minnesota
James Hodges, University of Minnesota

Social scientists have frequently attempted to decompose temporal trends in various outcomes into three aspects of time processes: age, period, and cohort. The problem that has faced researchers for decades is that the three distinct time processes are linearly related to each other (cohort=period-age), so disaggregation of temporal trends has to rely on problematic statistical assumptions. We develop a new method, called the age-period-cohort-interaction (APC-I) model, for analyzing age, period, and cohort effects. Compared with other age-period-cohort methods, the APC-I model has two advantages: First, it does not rely on problematic assumptions. Second, while other methods assume that cohort effects are constant from birth to death, the new APC-I model relaxes this assumption and allows researchers to examine changes within cohorts. Using 1974 to 2012 data from the General Social Survey, we demonstrate how this new model can be used to investigate inter- and intra-cohort variation in Americans’ political views.

  See paper

Presented in Session 214: Statistical Methods for Demographic Analysis