Further development of methodology for multivariate analysis of the total fertility rate and its components based on birth-history data

Abstract

A discrete-time survival model (the complementary log-log model) is used to model parity progression from woman's own birth to first marriage, from first marriage to first birth, from first birth to second birth, and so on, with one model for each parity transition. Predictor variables in each model include woman's age and duration in parity as well as socioeconomic characteristics. The models are applied to birth history data. Collectively the models yield estimates of marriage and birth probabilities by age, parity, and duration in parity, denoted Pait, by socioeconomic characteristics. The probabilities Pait are multivariate in the sense that they can be tabulated by one socioeconomic characteristic while holding other socioeconomic characteristics constant. The probabilities Pait allow construction of a life table that follows women by age, parity, and duration in parity from age 10 to age 50 one year at a time. This life table is referred to as a "global life table," because it spans all ages, parities, and durations in parity within the reproductive age span. Because the Pait are multivariate, the global life table is also multivariate, as are all measures derived from it. The derived measures considered here include both period and cohort estimates of parity progression ratios (PPRs), age-specific fertility rates (ASFRs), mean and median ages at first marriage, mean and median closed birth intervals, mean and median ages at childbearing (both overall and by child's birth order), total fertility rate (TFR), and total marital fertility rate (TMFR). The methodology is tested on data from the 2003 Demographic and Health Survey of the Philippines.