Strategies to Better Identify Student Homelessness Using Data in an Urban School District

• In Event: Students Experiencing Homelessness: Key Findings for Practice and Policy Decision Making

Abstract

Objectives
This paper tests strategies using LEA data to identify student homelessness. It also uses integrated data from other sources to better estimate its prevalence and inform multiservice partnerships required to for a comprehensive systems-response. The approach is the product of a research-practice partnership.

Perspective
Student homelessness threatens functioning, achievement, and attainment. Despite mandates to proactively identify and serve students experiencing homelessness, districts struggle because few have effective strategies or a systematic response to address students’ needs.

Methods
We evaluated 6 data-based strategies to identify students who meet the definition of experiencing homelessness. Strategies were suggested by the literature and through research-practice partnership activities. Chi-square goodness of fit tests evaluated whether each strategy increases the number of students recognized as experiencing homelessness relative to the number routinely identified by the partnering urban school district (See Table 1.1). One-sample t-tests evaluate attendance rate of students identified by each strategy relative to the district average (Figure 1.1).

Data sources
Data span 5 school years (2014-15 through 2018-19) and include routinely collected education records for the population of students in district-operated schools integrated using common identifiers at the individual level with municipal, health, and human service administrative records for students and their guardians.
Education data contained information on family relationships, address, LEA homelessness identification, and attendance rate (days attended / days enrolled).
Addresses of emergency shelter and hotels/motels were publicly available.
Health records included dates and relevant diagnoses for all hospital-based health care in the region. Contact notes were available for encounters in calendar years 2018 and 2019. These were subjected to text search with multiple terms indicating homelessness.
A property vacancy survey provided indicators for each parcel in the district geography, identifying addresses of properties in severe disrepair denoting substandard housing.
The municipal Department of Licenses and Inspections provided records for all addresses ordered to be boarded-up, indicating substandard housing.

Results
Each strategy identified a considerable number of students likely to have experienced homelessness but not identified by the district (min: 26; max: 109). Chi-square goodness-of-fit tests confirmed that each significantly increased the number identified (ps < .001). One-sample t-tests confirmed students identified through each strategy had lower average attendance than the district average (ps < .05).

Significance
These strategies suggest that student homelessness is more prevalent than recognized. Two use LEA data and public information only, making them appropriate for district outreach operations. Other strategies underscore the value of partnerships with municipal agencies and healthcare providers to better identify and serve students who may be living in substandard housing or other contexts of homelessness known to these systems. Policy and practice implications will be discussed.

Authors

• J. J. Cutuli, Nemours Children's Health System

• Sandra Torres Suarez, Camden City School District

• Aaron Truchil, Camden Coalition of Healthcare Providers

• Tyler Yost

• Ciani Green, Camden City School District