![[ Visit Client Website ]](images/banner.gif)
Accurately estimating the number of youth facing homelessness is crucial for effective federal and local policy-making as well as resource allocation. Under contract to the U.S. Department of Housing and Urban Development (HUD), we examined the capacity of regional administrative data to improve estimation of prevalence and incidence of youth experiencing homelessness in an urban county. The project leveraged an existing regional integrated data system holding individually-identifiable data from 35 public and private entities, the Child-Household Integrated Longitudinal Data (CHILD) system. The study’s main objectives were:
- To investigate the extent to which integrated administrative data systems can offer additional value to the task of estimating homeless youth, identifying the specific types of administrative data sources that provide the most benefit in counting and describing the homeless youth population;
- Develop a model for replication purposes in other communities, offering guidance to those just starting to coordinate data sharing agreements among partner agencies, as well as to those with already developed integrated data systems informing social and public policy in their communities.
Methods
In compliance with HUD’s request, we focused on youth aged 13 to 25 years old. We started by identifying youth receiving homeless services as seen in the Homeless Management Information System (HMIS) and school data. Using a novel address-list method we identified youth seen in CHILD administrative data, including food assistance data, to produce a registry of youth experiencing housing instability or homelessness from 2017-2019. We used this registry to estimate the number of youth not seen in the data using Multiple Systems Estimation (MSE) and contextualized our results with qualitative analysis from nine key informant interviews with thirty-one provider staff.
Results
The address list method proved valuable in identifying youth who do not connect with shelter services, yet might receive food assistance or interact with other systems. Applied to the resulting registry, MSE produced estimates roughly 2-3 times the amount of what is represented in the community administrative data. Results will be presented by age, sex assigned at birth, and racial strata used to produce them.
Key informant interviews identified themes related to nuances in administrative data, difficulty defining housing instability and homelessness, importance of program knowledge and lived experience to data interpretation.
Conclusions and Implications
Integrated data systems that leverage community administrative data provide crucial insights into the possibilities and limitations of enumerating youth experiencing housing instability and homelessness. Key informant interviews provide critical linkage of experiential and programmatic knowledge that guides data analysis and interpretation. MSE can be a useful tool for communities to better understand their local population of youth experiencing housing instability and homelessness but has statistical and skill limitations. Social work researchers play an important role in facilitating conversations between statisticians, administrative data staff, service providers, and youth with lived experience. These connections are of the utmost importance when considering the policy and funding implications of enumerating the population of youth experiencing housing instability and homelessness.