Making the Invisible Visible: Challenges in Estimating the Homeless Population

Arianna Gatta, The University of Queensland

In the last 40 years social scientists developed and applied a wide variety of methods to detect the homeless population. However, estimates tend to differ significantly depending on the method used. This leaves the question “How many homeless are there?” a difficult one to answer. In this work strengths and drawbacks of different quantitative methods are outlined. The city of Rome is used as case study, as the total homeless population has been estimated repeatedly over time using capture-recapture method (in years 2002-2011), S-Night count (in 2014), time-location sampling (in 2011 and 2014) and through the administrative data. Previous estimates are compared to those from a novel data collection I conducted in March 2021 in Rome using time-location sampling. Results show that homeless population estimates are highly sensitive to the contingent context of the Covid-19 pandemic. This calls for more attention when producing, using and comparing homeless estimates.

Keywords: Population projections, forecasts, and estimations, Small area estimation, Census data, Data and Methods

See paper.