Measurement Errors in Census Counts and Estimates of Intercensal Net Migration

Abstract

We present a method for generating confidence intervals around estimates of intercensal net migration, made using the life table survival method, that incorporate estimates of census measurement errors. The life table survival method applies a life table to a census count to project survivors at some past or future time points. Net migration is then estimated as the difference between the projected number of survivors and the enumerated population at that time. We present confidence intervals based on mean square error, the sum of the variance and squared bias. We assume that random variation in the number of net migrants in an age-sex group is due to random variation in mortality rates and to measurement errors in census counts. We illustrate the technique using data from a small area in Alaska.