A survey of software for fitting capture–recapture models

Abstract Capture–recapture analysis, also called mark‐ or multiple‐recapture, is aimed primarily at estimating the total size of a population. The population of interest may consist of animals, people, errors in complex software, the number of crimes committed by an oppressive political regime, coins struck by ancient dies, and so on. Statistical methods for population size estimation are well‐developed, with many extensions and variations such as allowing for birth, death or migration in the population; incorporation of predictor variables or spatial location of captures; observation by different physical methods, and so on. Accordingly, many software programs have been written and disseminated to implement these analyses, and a survey of those programs is given here. We classify the programs based on three different perspectives: types of classical closed‐population models, statistical foundations or philosophy, and extensions or variations of classical models. While the level of computing in this area has become quite sophisticated, especially for the extended models, none of the major statistical software packages has a ‘native’ capture–recapture sub‐package or routine (although some workarounds are possible), and the large number of separately released programs, though effective within their domains, tend to lack standardization and interoperability at present. The applied scientist can be reasonably confident of finding a program to fit his/her needs, but some examination of the literature will be required. WIREs Comput Stat 2013, 5:114–120. doi: 10.1002/wics.1250 This article is categorized under: Statistical and Graphical Methods of Data Analysis > Sampling