Estimating population sizes for capture–recapture sampling with binomial mixtures

Abstract

A binomial mixture model has been used to account for heterogeneous individual capture probabilities in capture–recapture studies. The problem of estimating the population size is reduced to estimating the odds that an individual is not captured. The odds is nonidentifiable. There is no upper bound to the odds. The odds admits lower bounds that can yield estimable lower bounds to the population size. An example is studied. A simulation experiment is used to evaluate the proposed and existing estimators.