“We do our work through adopting a compassionate theory of change in our governance, operation, and program delivery that’s based on engaging people from all walks of life to target sustained impact.”
Director of Research & Evaluation, SARAH
Homelessness affects people in every state, from urban to rural areas. On a single night in 2020, over 580,000 people in the United States were experiencing homelessness.
But while housing instability is a struggle faced by people all across the country, not everyone is equally at risk. Some groups are disproportionately represented in the homeless population, such as Black Americans, who make up 41% of those experiencing homelessness but only 13% of the U.S. population. Understanding exactly who is homeless and the implications of racial and other types of disproportionality are important considerations when striving to target aid to those most in need.
In 2021, staff at the South Alamo Regional Alliance for the Homeless (SARAH)--HUD’s lead agency coordinating a network of service organizations that aim to prevent and end homelessness in the San Antonio, Texas region--noticed a similar race discrepancy in their service area. Their annual point-in-time count--a method used nationwide to assess the state of homelessness in particular communities--revealed that the percentage of Black people experiencing homelessness was more than two and a half times that of Black San Antonio residents overall.
Compounding their concerns were questions about the effectiveness of the Vulnerability Index - Service Prioritization Decision Assistance Tool, or VI-SPDAT--an evaluation tool commonly used to measure the level of vulnerability of individuals experiencing homelessness and match them to appropriate services. The VI-SPDAT questionnaire asks people seeking housing assistance to report their age and other demographics and to recount their housing history and other markers of stability. A detailed point system pegged to response options calculates the respondent’s risk level when it comes to homelessness and other dangers. But the limited study of the VI-SPDAT finds that its ability to predict those most at risk of returning to homelessness is uncertain.
The SARAH team wondered, since so many Black community members weren’t getting the supportive services they needed to thrive, if others were likely also falling through the cracks, needs unseen and unaddressed. Could a new tool--one that better accounts for the unique factors and experiences of people of color and other historically underserved demographic groups--help ensure that housing assistance is more fairly and efficiently allocated to those in need?
SARAH uses an area-wide system known as Homelink to prioritize people for and direct them to homelessness services. Homelink relies on results from the VI-SPDAT to help service providers allocate resources in a logical, targeted way.
SARAH is addressing the VI-SPDAT’s possible shortcomings and contributions to disproportionate representation in the homeless community by developing a new tool to more accurately assess a person’s risk of homelessness and quickly match them with needed services. The new tool is much simpler and was developed specifically to make access to services more equitable.
As part of its core commitment to equity, SARAH is integrating the tool with historical data so it will more accurately assess the needs of populations such as Black Americans, who have been historically overrepresented in homelessness rates
What impact does a new assessment tool for prioritization that emphasizes various vulnerability indicators have on permanent supportive housing placement outcomes and equity?
The evaluation of SARAH’s new prioritization tool is a randomized controlled trial. When an individual comes to any of the agencies in SARAH’s network looking for housing assistance, they are screened for eligibility for services. If they are eligible and consent to be part of the research study, they are entered into a random lottery to determine if their needs will be assessed using the traditional VI-SPDAT or the new prioritization tool developed by SARAH. Those who are randomly chosen to be assessed by the new tool become part of the treatment group. Those randomly chosen to be assessed by the VI-SPDAT become part of the control group. The results of each person’s assessment is used to prioritize them for and connect them to services.
At the end of the study, LEO researchers will compare outcomes for the two groups to determine the impact of SARAH’s new prioritization tool on racial equity in service delivery and housing outcomes.
Impact Evaluation
Precisely measures the impact of a specific program by comparing the outcomes of two groups that are the same on average, except that one participated in the program and the other didn’t.
Randomized controlled trial
A rigorous form of impact evaluation that measures the impact of a program by comparing the outcomes of two randomly assigned groups that are the same, on average. Individuals in the treatment group participate in the program while those in the control group do not.
Quasi-experimental study
A form of impact evaluation that uses a tool other than random assignment--such as a program’s eligibility cut-off--to create two comparison groups and measure a program’s impact.
Retrospective study
Measures the impact of a past program by comparing the existing outcomes of a group of people who participated in a program with a similar group that did not. Relies heavily on administrative datasets, such as medical records or earnings data.
Natural experiment
An impact evaluation where participants are “naturally” sorted into treatment and control groups. For example, geography might decide whether or not an individual is impacted by a new law, and thus how the law impacts their outcomes.
Difference-in-differences
A quasi-experimental design that’s used when it’s not possible to randomize individuals into treatment and control groups. Before the intervention, researchers observe two groups of people that have the same concurrent outcome trends. After implementing an intervention that affects only one group, they compare the differences in outcomes across both groups.
3060-I Jenkins Nanovic Hall
Notre Dame , IN 46556 USA leo@nd.edu
