Building Evidence in the Fight to End Homelessness

This post is written by Rohit Naimpally, Senior Research and Policy Manager at J-PAL North America, a regional office of the Abdul Latif Jameel Poverty Action Lab (J‑PAL). J-PAL is a global research center based at the Massachusetts Institute of Technology (MIT) working to reduce poverty by ensuring that policy is informed by scientific evidence.

For decades, organizations across the United States have been running innovative programs to reduce homelessness, prevent evictions, and promote housing stability—almost always with limited resources.

Now, more than ever, we must prioritize the evidence behind our efforts. Homelessness has tragic impacts on the individuals and communities affected by it, so it is essential that the programs we choose to combat it are rooted in credible research.

Rigorously evaluating homelessness prevention and reduction programs through initiatives like the J-PAL North America Evaluation Incubator can help service providers improve their programs and policies, gain buy-in and support scale-up, and understand what is working and why.

Designing strong impact evaluations related to homelessness

While more and more communities are investing in efforts to respond to the homelessness crisis, key questions remain on which strategies are most effective in addressing housing instability.

For example, among households at risk of homelessness, is it helpful to provide individualized assistance from a case manager to improve housing stability, or is the provision of emergency financial assistance enough? Does rapid re-housing help individuals overcome barriers to long-term housing stability, or is the assistance provided too limited to have lasting impacts?

Housing instability is shaped by numerous structural and contextual factors. The complexity of the problem necessitates that we rigorously test programs aimed at addressing housing instability to better target assistance to unhoused individuals and prevent homelessness before it even occurs.

Evaluating programs and using data to understand and improve them can take many useful forms. However, randomized evaluations are generally considered to be the strongest research design to measure the average impact of a program.

In a randomized evaluation, eligible participants are randomly sorted into two groups—one that is offered access to a new program and one that receives standard care—ensuring that at the beginning of the study the groups will be the same on average. When properly designed and implemented, these studies are uniquely positioned to enhance our understanding of which strategies are most effective for reducing homelessness and can inform policy action.

Critically, a well-designed evaluation will not only demonstrate whether a program is working to address homelessness, but will also speak to why it is working. For example, an ongoing randomized evaluation of a youth and family homelessness prevention initiative in King County, Washington is measuring the effectiveness of a combined progressive case management and flexible financial assistance program compared to the provision of only flexible financial assistance. This will allow local lawmakers to isolate the impact of the program’s case management component—answering questions for researchers and decisionmakers alike about the relative importance of case management in financial assistance programs to prevent homelessness.

Using high-quality research results to advocate for policy changes

A rigorous evaluation can help to generate buy-in and increase support for a program or policy.

Before the early 2000s, unhoused individuals in the United States typically had to meet tests of “housing readiness” before progressing from shelter to transitional housing. Several randomized evaluations of Housing First—which prioritizes the provision of permanent housing with no preconditions—found that individuals in permanent supportive housing spent about half as much time unhoused or in the hospital. These results flipped the conventional logic of homelessness services on its head and supported the adoption of a Housing First approach across the country.

By demonstrating what approaches have been successful, randomized evaluations can have an impact far beyond the original context in which they were conducted. Rigorous evidence equips implementing organizations with an understanding of which approaches have worked elsewhere and allows policymakers to expand programs with demonstrated success. In Santa Clara, California, an ongoing randomized evaluation testing the impacts of a rapid re-housing intervention for single adults experiencing homelessness will directly inform the County’s decision to scale the program. Beyond this, evaluation results will be disseminated to policymakers in other regions to inform the design and replication of future housing initiatives.

Overcoming barriers to using and generating evidence

Organizations often face a number of barriers to running randomized evaluations:

  • They can demand significant staff time, which can be particularly challenging for overburdened practitioners.
  • They can require an up-front investment of time and resources, while many of the benefits are not realized until a much later date.
  • They can be difficult to integrate into day-to-day program operations.
  • And while many organizations are required to demonstrate evidence-based practices to receive funding, existing funding sources are often dedicated to program delivery and program implementation rather than to evaluation—creating a catch-22 for those that wish to evaluate to receive resources but require resources in order to do so.

The pathway forward necessitates that practitioners, researchers, and policymakers collaborate to run high-quality impact evaluations and expand the base of evidence on strategies to reduce and prevent homelessness. Yet, it is clear that many practitioners require technical and financial assistance to do so.

J-PAL North America has developed a Housing Stability Evaluation Incubator to support organizations working to end homelessness in using randomized evaluations to generate new lessons through rigorous research. Through the Evaluation Incubator, selected organizations will receive pro-bono technical assistance from J-PAL’s staff of policy and research experts, flexible funding, and connections with leading academic researchers to develop a high-quality evaluation.


To learn more about this opportunity, including instructions on how to apply, please visit the J-PAL’s Housing Stability Incubator web page. Preliminary letters of interest are due May 4, 2020.

While addressing homelessness in the United States can seem like an insurmountable challenge, we have powerful tools for learning more about which strategies are working and why. Organizations across the country have been developing creative solutions to housing instability for decades, and with rigorous research, can answer critical questions about how to best target limited resources, address unique challenges in their local contexts, and improve the nation’s policy response to the homelessness crisis.