Estimating the Influence of Fixed Covariates on Long-Term Survival Using Repeated Cross-Sectional Data

Scott M. Lynch, Princeton University

Studies since the late 1970s have shown how differential rates of mortality of members in a cohort affect the composition of the cohort with respect to fixed characteristics. For example, although males outnumber females at birth, higher mortality rates among males produce a cohort that becomes increasingly and predominantly female across age. Although research has often focused only on how mortality affects a cohort’s composition, changes in a cohort’s composition can be used to estimate the influence of fixed characteristics on survival, and repeated cross-sectional data are all that are required. Others have developed methods that are limited to two cross sections and/or can only estimate the influence of fixed characteristics on relative survival differences. In this paper, I develop two methods for estimating the influence of fixed covariates on absolute (not relative) survival using multiple waves of repeated cross-sectional data from the 1972-2008 General Social Survey.

  See extended abstract

Presented in Session 17: Innovations in Measurement and Modeling