Identification and Estimation of Age-Period-Cohort Models in the Analysis of Discrete Archival Data

We formulate an age-period-cohort specification for discrete data.When ages and periods are evenly and equally spaced there is an identification problem .We explain the problem.Sampling models are indicated and the minimal sufficient statistics are given.The iterative proportional fitting and Newton-Raphson methods for maximum likelihood estimation are described for age-period-cohort specifications.Various formulations of discrete response variables are listed, and their special features explained.We also consider the identification problem when I