Parametric regression models for continuous time removal and recapture studies

• Published in: Journal of the Royal Statistical Society. Series B, Statistical methodology Vol. 61; no. 2; pp. 401 - 411
• Main Authors: Lin, D. Y., Yip, P. S. F.
• Format: Journal Article
• Language: English
• Published: Oxford, UK and Boston, USA Blackwell Publishers Ltd 01.01.1999; Blackwell Publishers; Royal Statistical Society
• Series: Journal of the Royal Statistical Society Series B
• Subjects: Animal abundance; Aviculture; Bird banding; Capture-recapture experiment; Consistent estimators; Counting process; Estimation; Estimation methods; Estimators; Heterogeneous capturability; Inference; Martingale; Monte Carlo simulation; Parametric models; Population estimates; Population size; Population size estimation; Regression analysis; Reliability testing; Standard error; Statistical analysis; Statistical models; Statistics
• DOI: 10.1111/1467-9868.00184
• ISSN: 1369-7412

We use a class of parametric counting process regression models that are commonly employed in the analysis of failure time data to formulate the subject-specific capture probabilities for removal and recapture studies conducted in continuous time. We estimate the regression parameters by modifying the conventional likelihood score function for left-truncated and right-censored data to accommodate an unknown population size and missing covariates on uncaptured subjects, and we subsequently estimate the population size by a martingale-based estimating function. The resultant estimators for the regression parameters and population size are consistent and asymptotically normal under appropriate regularity conditions. We assess the small sample properties of the proposed estimators through Monte Carlo simulation and we present an application to a bird banding exercise.