Incorporating cyclical effects and time-varying covariates in models for single-source capture-recapture data

Abstract

The objective of capture-recapture analysis is to estimate the size of an elusive population, for which the zero-truncated Poisson model is a basic model. We extend this model to the more general recurrent events model to include cyclical eects and time-varying covariates. An application to police data on victims of domestic violence provides strong evidence for the presence of weekly and seasonal cyclical eects on the rate of police reports.