Detail from "the second line," a painting by Bob Graham. For more about the artist, click here.

Online Program

Fixing National Homeless Counts

*Jay Bainbridge, Marist College School of Management 

Keywords: measuring undercounts for HTR groups, sampling HTR populations, homeless and refugee populations

Since 2003, the New York City Department of Homelessness has conducted an annual Homeless Outreach Population Estimate (HOPE), which surveys the unsheltered homeless individuals residing in public spaces in New York City. Volunteers led by experienced team leaders canvass a random sample of above ground areas comprising a few city blocks each, and subways stations and trains, and ask everyone they encounter a short series of questions to determine whether they may be homeless; the results are used to estimate the total number of homeless persons per borough and in the subways.

NYC’s efforts were inspired by the Rough Sleepers Initiative in Central London, and preceded national mandates by the U.S. Department of Housing and Urban Development (HUD) directing all localities to conduct at least biennial counts. NYC’s approach has since been recognized by HUD as a gold standard for statistically valid methodologies. In addition, NYC is unique in being the only known U.S. municipality to use plant-capture techniques to adjust for under-counts, by hiring an outside evaluator to place decoys in randomly selected survey areas.

This paper reflects briefly on the practical strengths and weaknesses of the methods used, including proposals for keeping a consistent measuring stick in light of a potentially growing sample size with limited volunteer resources. Several strategies for managing the operational challenges to conducting on-going large-scale survey are discussed. Along with analysis of the decoy initiative, the paper also offers a model for homeless outreach workers to incorporate random surveying into their regular work, to improve their knowledge of a changing population and to provide more than a singular (and potentially largely varying) data point to gauge the success of their work. These findings have implications for other localities looking to improve their surveys of, and services for, the homeless.

ASA Meetings Department · 732 North Washington Street, Alexandria, VA 22314 · (703) 684-1221

Copyright © American Statistical Association.