Communities across Texas are gearing up for their annual homeless count in January. Homeless counts are a tradition in many communities, and they are required at least every two years for all communities applying for the Department of Housing and Urban Development’s (HUD) annual Continuum of Care (CoC) application. This year’s count will be the first time HUD will not allow the use of estimates in the official report of the homeless count, which troubles many communities that fear they will be dramatically under-representing their homeless population without the use of some estimation.
HUD has been trying to improve the accuracy of homeless counts over the past several years. The CoC application requires communities to conduct a point-in-time (PIT) count to reflect the level of homelessness at a particular moment in time. Some communities conduct their counts over several days or even weeks, which is acceptable as long as they collect information about one specific day, usually the night before the count begins.
Since the 2003 CoC application, HUD has requested that the counts take place in the last week of January and has required communities to collect local data on the number and characteristics of unsheltered homeless people, specifically the chronically homeless. The PIT count must include the number of homeless people sleeping on count night in emergency shelters, transitional housing, and their equivalents (e.g., motels or hotels on “homeless” vouchers) and the number of non-sheltered homeless people, which requires a great deal of planning a very large number of volunteers. The Tarrant County coalition reports that they have enlisted over 500 volunteers to count the unsheltered homeless for their upcoming count night on Jan. 26, 2006.
The biggest challenges facing any community planning to conduct a PIT count are the coverage (whether they will be able to include all the relevant people) and duplication (finding a way to ensure no one is counted more than once). Reporting on the number of homeless individuals “in” shelters and accessing homeless services is fairly straightforward and has actually been made easier with the use of Homeless Management Information System (HMIS), the required tracking system used by all HUD-funded continuums of care. A point-in-time, unduplicated count can be produced with a report for any day of the year. Not all homeless service providers within a continuum of care are a part of the HMIS reporting, so the survey of these additional service providers is generally straight forward.
The unsheltered count is far more difficult to conduct and to get accurate numbers. Most communities do their best to conduct counts of local homeless populations based on sound methodology, but most are limited by a lack of money and time. Too many rely on last-minute planning and volunteer recruiting.
There are two approaches that HUD offers to communities to conduct their unsheltered or “street” counts.
The first approach is to simply report the number of people counted, which is most often accomplished by a one-night “street” count. The problem with this approach is that is misses many homeless people. It is especially problematic in rural areas where roads are sometime impassable, and volunteers are not willing to go in by foot. Another population that is really hard to capture with a PIT count is families living in cars. A simple PIT count generally skews the data primarily toward single adults and unaccompanied youth, who are more likely to be “seen” on the streets during the one-night count.
The second approach endorsed by HUD is to use statistical sampling and extrapolation to arrive at an estimate of the number of unsheltered homeless persons. HUD does not want communities using adjustment or inflation factors that do not have any statistical basis. This had many communities concerned that their methodology will not hold up to HUD’s new standards. Several Texas communities have even decided to skip a count this year because of the restrictions on the use of estimates.
The Coalition for the Homeless of Houston/Harris County reported their actual PIT count numbers in their 2005 application but estimates that 30-35% of their homeless population was not included in the count. The City of Austin reports that their count in January, 2005 “reached only 38 encampments in the local area, yet staff from the ACCESS program of the Austin Travis County MHMR Center and Park Police officers from the City of Austin’s Parks and Recreation Department indicate that almost 400 encampments exist in Austin/Travis County; therefore, the PIT count only included 10% of the sites where persons are living in homeless encampments. With serious concerns about HUD’s increased emphasis on “statistically valid methodology,” the City of Austin has decided not to hold a count this year.
The methodology that HUD recommends in its publication A Guide to Counting Unsheltered Homeless People is the approach used in New York City, which utilized a system of “statistically valid” sampling based on neighborhoods classified as high to low density of homeless people expected to be found there. This methodology is extremely time consuming and costly. The Guide states, “the two staff members from New York’s Department of Homeless Services who organized New York City’s 2003 “street” count estimates that they spent 100 percent of their time preparing for the count in the month leading up to it. One staff member worked full-time the month after the count to complete the data analysis.”
The Guide further states that Washington’s Balance of State CoC spent approximately $4,500 on a technical assistance consultant and used staff time from the CoC lead agency valued at approximately $5,000. Atlanta’s 2003 count of sheltered and unsheltered populations cost approximately $120,000, and Long Beach, CA’s 2003 count had a final cost of approximately $150,000 (which also included an in-depth assessment and 10-year planning process).
What communities decide to do and how HUD will score the methodologies currently used will be important for Texas communities to watch in the next CoC funding cycle. HUD does offer some advice when planning your community’s next required count in 2007:
Don’t reinvent the wheel. Chances are other communities of your size, geography, and resources have conducted a count. Talk with them about how they have conducted their public places count. Consider using their approach as a blueprint for your first count.
Communicate with your community. Begin the process of discussing an unsheltered homeless count with members of your CoC and others in the community long before you begin the formal planning. Almost every community that conducts a successful public places count emphasizes the importance of participation and “buy-in” from different groups.
Be prepared for media attention. Not every count will receive media attention. However, it is important to think about the purpose of he count and the message you might want to convey to the media, if necessary. Several communities use their regular public places count to draw attention to the issue of homelessness
In our next issue of Homeless Network News, we’ll cover the survey methods used by Texas communities. If you have any comments to share, please email newsletter@thn.org
|
