Learning More From Homeless Point-in-Time Counts

Previous research has found that patterns of sheltered homelessness and characteristics of people who use shelters vary between point-prevalence and period-prevalence tallies, but there are no comparable data for unsheltered homelessness. We derive a three-year period-prevalence measure of unsheltered homelessness from data collected in a point-in-time count, by weighting individuals by the inverse number of months, out of the last 36, that they report being homeless. We illustrate using data from the 2022 Point-in-Time Count in Nashville/Davidson County Tennessee, which found 634 unsheltered people. We estimate that the three-year prevalence of unsheltered homelessness would be at least 3.5 times the one-night count, but that far fewer people would meet the US Department of Housing and Urban Development's definition of chronic homelessness. Over half of people sampled outdoors, especially those living with household members and pets, report never using shelter. Point-in-time counts can yield far more information useful for planning than the simple count itself.