The Influence of School Quality on the Impact of Assisted Housing in Milwaukee

• In Event: 25.033 - A Multisite Study of Assisted Housing Programs and Educational Well-Being Using Integrated Data Systems

Abstract

Our Milwaukee research team was able to access a larger portion of education records to create state-of-the-art measures of a school’s performance relative to other schools in the district. We were then able to test whether students in assisted housing programs were more likely to end up in higher performing schools compared to their matched peers, or whether the performance of the school children attended influenced the impact of assisted housing on other important educational outcomes.

We answered these questions using an advanced research design that improved upon the multi-site base research design. The base research design examined a single cohort of students starting in third grade and in seventh grade. Due to sample size concerns, this method did not allow for a measure of educational status prior to the application of assisted housing treatment. To alleviate this concern, the analysis of Milwaukee used a stacked cohort of four separate cohorts of students starting in third grade. By increasing the sample size, we were able to identify students joining the treatment during the years of interest and measure educational outcomes both before treatment and after treatment. The analysis model still included propensity score matching to create a control group. In this modified method, matching occurred within each of the four stacked cohorts. The analysis then differed from the base research design by using a difference in differences model with the matched sample. This modified model controlled not only for treatment but also for any differences that might occur between students who eventually participate and students who did not participate at all in assisted housing. A school performance metric was then created by calculating an average value-added estimate in mathematics and reading for each school. These value-added estimates examined the average growth attributable to a school on mathematics and reading test scores while controlling for student prior academic experience and demographic characteristics.

Results indicated that students in assisted housing programs were less likely than their matched peers to enroll in high-performing schools by the end of the study period. By employing the difference in differences model with school value-added estimates included, results indicated a small negative association between assisted housing and reading growth, no association between assisted housing and unexcused absences or graduation rate, and mixed findings for the relation between assisted housing and number of suspension days. These findings suggest that students in assisted housing could benefit from greater cross-agency collaboration between housing and education to ensure access to high-performing schools. This partnership holds potential for changing the educational trajectory of these students.

Authors

• Robert H. Meyer, Education Analtyics

• Grant Sim, Wisconsin Center for Education Research

• Kaveh Akram, Wisconsin Center for Education Research