Using local longitudinal records to estimate transient and resident homeless populations

Abstract

Many local agencies that provide services to the homeless have only limited data and resources to estimate the extent of need or the size of the local population at risk of requiring the services they provide. Methods exist, however, that are capable of providing reasonable estimates of underlying service populations with minimal record keeping. In this paper one group of such models, first‐capture models, is discussed as a resource‐efficient technique for estimating the underlying population consistently at risk ("residents") and those occasionally at risk ("transients") of requiring agency services.