Co-Principal Investigator: Franco, Ximena; Winsler, Adam

Purpose: In this project, researchers focus on exploring relationships between language of instruction and academic outcomes in the context of dual language (DL) educational programs. The research team will look at two-way immersion DL programs, which provide instruction in both English and a non-English language (in this case, Spanish) to both English learners (ELs) and English-speaking students. DL programs have been posited as a mechanism to reduce the achievement gap for ELs, but more research is needed about how and why such programs may support learning and achievement.

Project Activities: In this study, researchers will collect data from K–3 students in DL programs to explore whether the amount of instruction they receive in English and in Spanish predicts fall to spring gains in academic skills in English and Spanish. The researchers will collect data from two schools each during Years 1 and 2, and collect data from students and teachers in the fall and the spring. Student data will include language skills, academic outcomes, and academic self-concept; teacher data will include questionnaires about student learning problems, approaches to learning, study skills, and student-teacher relationships. The researchers will observe classrooms to measure student engagement and the amount of language (Spanish or English) used during instruction. In addition to predicting spring outcomes, the researchers also will explore whether any observed associations vary by student proficiency in each language, and whether they are mediated by student engagement, academic self-concept, approaches to learning, and student-teacher relationships.

Products: Researchers will produce preliminary evidence about the relationships between language of instruction and student academic outcomes, as well as moderators and mediators of these relationships. The researchers also will produce peer-reviewed publications.

Structured Abstract

Setting: The study will take place in four elementary schools in two North Carolina districts. One district is suburban, and the other is more rural.

Sample: The sample will include teachers from 50 DL classrooms and 300 randomly sampled students from those classrooms. Most classrooms are taught by two teachers — one who teaches in English, and one in Spanish — and both must participate to be included. All K – 3 students (ELs and non-ELs; approximately 1200 students in all) enrolled in the 50 classrooms will be invited to participate in the study, and the researchers will randomly select three volunteer students from each classroom each year (150 in Year 1 and 150 in Year 2). Random selection will be stratified by student home language (English, Spanish, or bilingual). The student populations in participating schools include sizeable numbers of students identified as English Learners (28–63%), Latino (32–65%), and low-income (38–90%).

Malleable Factors: The primary malleable factor in this study is language of instruction. Researchers will focus on the amount that each language is used in two different types of DL models. In the 50/50 model, students receive half of their instruction in English and half in Spanish, maintaining this division as they progress through the grades. In the 90/10 model, students begin kindergarten receiving 90% of their instruction in Spanish, and this percentage decreases incrementally in the subsequent grades until a 50/50 distribution is reached. For both models, the research team will explore relationships among language of instruction and a number of other malleable factors, including student engagement, academic self-concept, approaches to learning, student-teacher relationships, language proficiency, and academic outcomes such as achievement, retention, and report card grades.

Research Design and Methods: Researchers will collect data from two schools each during Years 1 and 2 (one school per district each year). Recruitment of teachers will take place during teacher pre-service meetings prior to the beginning of the school year, and students and families will be recruited during the first four weeks of the school year. Researchers will assess students in Spanish and English in the fall and the spring on measures of language skills, academic outcomes, and academic self-concept. They will observe students' classroom engagement twice during the winter. Teachers will complete questionnaires about students' learning problems (at fall and spring), approaches to learning, study skills, and student-teacher relationships (during winter). Observers and teachers will rate the amount of Spanish and English instruction in the fall and spring. The researchers also will work with school partners to organize the transmittal of administrative data on retention, report card grades, and background variables, at the end of each school year.

Control Condition: Given the exploratory nature of this grant, there is no control condition. Because the language of instruction will be modeled using a dosage framework, however, the implicit comparison for the central analyses is to conditions where students receive less, or no, instruction in their home language.

Key Measures: The research team will use instruments such as the LAS-Links, the Woodcock-Johnson IV, the Language Use Inventory, the Behavior Assessment Scale for Children, the Learning Behaviors Scale, and the Student-Teacher Relationship Scale. They will also administer surveys and conduct classroom observations.

Data Analytic Strategy: Researchers will conduct descriptive analyses of participating students, families, teachers, and classrooms. They will use multilevel linear models to examine the association between language of instruction and within-year change in academic outcomes, and whether this association is moderated by student language proficiency in English and in Spanish. The team also will examine the dimensionality of certain mediators, and use structural equation modeling, regression analyses, and hierarchical linear modeling to explore the relations among language of instruction, mediators, academic outcomes, and language proficiency.