Enumerating the Hidden Homeless: Strategies to Estimate the Homeless Gone Missing From a Point-in-Time Count

• Main Authors: Agans Robert P., Jefferson Malcolm T., Bowling James M., Zeng Donglin, Yang Jenny, Silverbush Mark
• Format: Article
• Language: English
• Published: Sciendo 2014-06-01
• Series: Journal of Official Statistics
• Subjects: homeless count; hidden homeless; unsheltered homeless population; horvitz-thompson estimator; multiplicity-based estimator
• Online Access: https://doi.org/10.2478/jos-2014-0014

To receive federal homeless funds, communities are required to produce statistically reliable, unduplicated counts or estimates of homeless persons in sheltered and unsheltered locations during a one-night period (within the last ten days of January) called a point-in-time (PIT) count. In Los Angeles, a general population telephone survey was implemented to estimate the number of unsheltered homeless adults who are hidden from view during the PIT count. Two estimation approaches were investigated: i) the number of homeless persons identified as living on private property, which employed a conventional household weight for the estimated total (Horvitz-Thompson approach); and ii) the number of homeless persons identified as living on a neighbor’s property, which employed an additional adjustment derived from the size of the neighborhood network to estimate the total (multiplicity-based approach). This article compares the results of these two methods and discusses the implications therein.