Bayesian Plant-Capture Methods for Estimating Population Size from Uncertain Plant Captures
Conference
65th ISI World Statistics Congress 2025
Format: IPS Abstract - WSC 2025
Keywords: bayesian hierarchical model, capture-recapture, homeless, missing-at-random
Session: IPS 733 - Bayesian Model Based Methods with Applications
Tuesday 7 October 2 p.m. - 3:40 p.m. (Europe/Amsterdam)
Abstract
Plant-capture is a capture-recapture method used to estimate the size of a population. In this method, decoys referred to as "plants" are introduced into the population. These plants are assumed to seamlessly blend with the population and are indistinguishable from regular individuals. The proportion of plants that are captured provides an estimate of capture probability. The inverse of this capture probability is then employed to extrapolate the population count into an estimate of the population size. However, existing plant-recapture methods typically do not consider the uncertainty in the capture status of the plants, particularly in the context of point-in-time surveys of the homeless. In this research, we introduce a range of Bayesian models designed to formally address this uncertainty. We validate the statistical performance of these models through simulation studies. Furthermore, we apply this novel methodology to estimate homeless population size in various U.S. cities in the context of the large-scale Shelter and Street-Night (S-night) survey conducted by the U.S. Census Bureau.