Dear Colleagues. A more nicely formatted version of this note, at www.barkerstats.com/PDFs/ASA/CNSL/Point-In-Time.pdf
excerpting
Preface
Count of the Unsheltered Homeless
First a disclaimer, these are entirely my own recommendations and my due diligence.
Most if not more Importantly, I appreciate perspectives from any Section or ASA member involved in any way whatsoever with "counting the homeless"
I prepared this for statisticians in the Section or any statistician interested in -considering- to volunteer for counting the Homeless, as part of a US-wide effort organized by HUD (Housing and Urban Development). The timing is that the (with exceptions ) entirely volunteer effort will occur in late January 2023. And those interested can begin to look now for their local "continuums of care" (C.O.C.) organizing the count. For example, I volunteered and my C.O.C. at the time was San Mateo County. If I were to volunteer again I would find the C.O.C. for Sacramento County (cited below). This note is my due diligence about the homeless count and as I outline below there may be significant caveats to know about the count before volunteering. And I think there is an opportunity for research projects to -prospectively- improve the survey- I'll defer to others as to the merits of that research.
The count is of the "unsheltered homeless" – definitions provided by HUD and cited below. This "count" occurs every other year, next occurring in 2023. The count was rescheduled for 2022 in Sacramento due to the pandemic. In 2011, I volunteered for this activity myself. I would consider participating in the survey again - however I offer at least one major Caveat Emptor. My caveat - there are some harsh criticisms of the so called "Point in Time" count of the homeless, and criticisms of the quality and accuracy of the data collected – nothing suggests these can't be improved. For example, one of if not the most-harsh criticism, is that simply because the count is conducted in January, a winter month the count is designed to "under count" the homeless. Below I provide links to my 2011 blog entry about my one-time homeless count experience. And I include further background which I undertook for my due diligence . That includes reference to some of the statistical literature (the article from the Annals of Statistics , possibly foremost) on potential statistical methodological issues. I provide a citation to methodology research ideas from UMN- School of Public Health (under Hennepin County). There have been some evaluations of the accuracy and quality by methods such as "capture recapture" and "decoy methods".
In my experience in San Mateo county, described in my blog entry, there was no direct confirmation that a homeless person was in fact homeless, the count of the homeless was entirely by observation from a distance and we were instructed not to directly interact with homeless for "safety reasons". The city of Oakland, in Alameda County uses Homeless as "Guides" to finding the homeless and thereby may also use a "no direct contact" count of the homeless but "augmented" the count with interviews of the homeless (https://everyonehome.org/wp-content/uploads/2022/10/2022-Alameda-County-PIT-Report_9.22.22-FINAL-3.pdf ).
------------------------------
Chris Barker, Ph.D.
2022 Statistical Consulting Section
Chair-elect
Consultant and
Adjunct Associate Professor of Biostatistics
www.barkerstats.com---
"In composition you have all the time you want to decide what to say in 15 seconds, in improvisation you have 15 seconds."
-Steve Lacy
------------------------------