Southwest
Description: Many states, including Texas, categorize schools as "low performing" based on student scores on standardized tests, sometimes along with low graduation rates and other achievement indicators. A primary concern of Texas stakeholders is how to identify, monitor, and support persistently low-performing schools in the next few years in the context of no STAAR data in 2020 and STAAR data in 2021 that may be less representative or reliable than in prior years. REL Southwest researchers will assist the Texas Education Agency by (1) exploring relationships between potential predictors not currently used for school performance ratings (including student attendance rates, course grades, disciplinary records, and educator characteristics) and school performance ratings, (2) developing binary indicators based on school-level predictors that are significantly associated with school accountability ratings. These indicators are particularly useful to identify schools most in need of supports when accountability ratings are not available or when their reliability is uncertain. The Texas Education Agency can use the indicators to identify schools that may be at risk of low performance and to inform decisionmaking about the allocation of resources and supports for school improvement.
Research Questions:
Study Design: REL Southwest will address the research questions by conducting school-level analyses. First, the study team will extract school-level files from the Texas Education Agency's website and obtain deidentified administrative records on students attending public schools in Texas between 2017/18 and 2020/21. Next, the study team will conduct regression analyses to determine which potential predictors of course enrollment and completion, attendance, or disciplinary actions, or educator characteristics are associated with 2017/18 and 2018/19 Texas school performance. Using the identified predictors, the study team will create a set of binary indicators that accurately predict whether schools met accountability expectations. Finally, the study team will apply the indicators to Texas public schools in 2019/20 and 2020/21 to determine whether identifications of low-performing schools are reliable and consistent.
The analyses will be conducted separately for elementary schools, middle schools, and high schools. The researchers anticipate that the study will be more likely to identify useful indicators for schools that include high school grades because of the larger number of non-STAAR potential predictors available.
Projected Release Date: October 2022
Related Products: This project will result in two products: a 15-page research report and a 1-page summary of findings.
Yibing Li, American Institutes for Research
Marshall Garland, Gibson Consulting