Co-Principal Investigator: De La Torre, Maria

Purpose: This study will systematically examine the experiences and performance of high school English Learners (ELs), alert to differences among distinct subgroups: long-time ELs, former ELs, and late-arriving ELs, to help schools better support their educational attainment. ELs are less likely to graduate from high and enroll in college, but little is known about the factors most important for success among high school ELs.

The project has four goals: 1) identify EL-specific early warning and college readiness indicators; 2) provide needed information on EL achievement on metrics other than test scores and attainment, paying attention to differences among subgroups of ELs (including long-term, late-arriving, and former ELs); 3) identify malleable school and district factors that have the highest potential leverage for influencing ELs' academic performance in high school and educational attainment; and 4) identify the ways in which the switch to the COVID-19 pandemic and switch to remote learning influenced the educational attainment and high school experiences of different groups of ELs, and which school practices supported stronger outcomes.

Project Activities: The University of Chicago Consortium on School Research (Consortium), with the Latino Policy Forum (Forum), and the Chicago Public Schools (CPS) Office of Language and Culture (OLCE), will conduct an exploratory study of the educational attainment and achievement outcomes of ELs in CPS, one of the top 10 EL-serving districts in the country. The project is possible because of an extensive data archive that links multiple distinct data sets through common student and school identifiers, from kindergarten through college graduation. They include student background, English proficiency, course taking, grades, attendance, discipline, student reports of their school experiences, teacher certification, college enrollment and attainment. Using ten freshmen cohorts (class of 2012-class 2021), Consortium researchers will conduct descriptive analyses, logistic regression models, models with school fixed-effects, and hierarchical generalized linear models. The Forum will convene an advisory group and work with researchers on the production of policy- and community-friendly products and dissemination activities. CPS-OLCE officials will meet regularly with Consortium and Forum project team members to ensure the research and products are useful for schools and districts.

Products: The proposed products include academic articles on early warning and college readiness indicators for subgroups of ELs and malleable school factors that look to have the most leverage for improving ELs' educational attainment, as well as briefs, one-pagers, and/or infographics targeted for non-academic audiences, guided by the advisory group and CPS-OLCE.

Structured Abstract

Setting: The setting for this project is Chicago Public Schools (CPS), Chicago, Illinois. CPS is an urban district of about 350,000 students, with 63,000 active English Learners (ELs) in the 2020–21 school year.

Sample: This project will focus on a population of students entering ninth grade in CPS in fall 2008 through fall 2017, with data linked from their elementary grades through their post-secondary years. Each cohort has approximately 28,000 first-time ninth graders, of which between 2,000–3,000 were active ELs in high school, and 6,000–9,000 were students formerly classified as ELs. There were 173 schools serving ninth graders in the fall of 2017 and 118 schools serving ninth-grade active ELs. The overall composition of ninth-grade classes was about 81% free-reduced price lunch (FRPL) eligible, 16% with an Individual Education Program, 48% Latino, 37% Black, 9% White, and 4% Asian. Among active ELs in ninth grade, 86% received FRPL, 37% had an IEP, 83% were Latino and spoke Spanish at home, and 70 different languages were represented.

Factors: The study develops early warning and college readiness indicators for ELs as a group, and subgroups of ELs (long-term, late-arriving, and former ELs), and then identifies the ways in which malleable school and district factors influence students' experiences and achievement in high school, and their eventual educational attainment. Malleable school and district factors include students' coursework (e.g., AP/IB/honors, STEM, college-prep English, ESL, remedial, arts), course experiences (e.g., rigor, clarity, personal support), isolation/integration (% EL and % same language in classes), teacher certifications (bilingual, special education), school discipline practices (e.g., use of in-school and out-of-school suspensions), school climate (safety, academic focus, college going culture, parent-teacher trust — both overall and based on EL students' reports), accountability ratings (which could inform district school choice systems), and college feeder patterns (which could inform postsecondary advising in high schools).

Research Design and Methods: The team will analyze outcomes and experiences of ten different cohorts of ninth-grade students, following them from ninth grade to the end of high school. Two cohorts will be left out to run confirmatory analyses on the final models.

Key Measures: The primary outcomes are indicators of educational attainment: high school graduation, college enrollment, college persistence, and college degree attainment. The study also examines mediating academic achievement outcomes: course grades, test scores and gains, English proficiency levels, and English reclassification. To understand students' experiences in high school as mediators of school structures on academic achievement, the study examines students' perceptions of school (school connectedness, relationships with peers, emotional health, feelings of safety, parent supportiveness in school), their academic behaviors (attendance, suspensions, study habits, and grit), and their course experiences (types, composition, and quality).

Data Analytic Strategy: The team will use a combination of descriptive statistics, logistic regression models, school fixed-effects models, and hierarchical generalized linear models. They will use statistical models to control for potential spurious factors and selection effects when making conclusions about the potential influence of particular school or classroom factors through baseline indicators of achievement in middle school, and school and student fixed-effects.