|Title:||A Multisite Randomized Controlled Trial (RCT) of Descubriendo la Lectura (DLL) Topic Area: English learners|
|Principal Investigator:||Borman, Trisha||Awardee:||American Institutes for Research (AIR)|
|Program:||English Learners [Program Details]|
|Award Period:||4 years (7/1/2016-6/30/2020)||Award Amount:||$2,918,455|
|Type:||Efficacy and Replication||Award Number:||R305A160060|
Purpose: The purpose of this study is to evaluate the efficacy of the Descubriendo la Lectura (DLL) program, which is the Spanish reconstruction of Reading Recovery, an English literacy program. The population of Spanish-speaking English learners (ELs) in United States schools is large and growing, and many of these students choose to enroll in bilingual instructional programs in early grades. As the number of students receiving initial literacy instruction in Spanish increases, the need for a Spanish language early intervention program for students at risk of literacy problems also increases. DLL has been evaluated previously using a randomized controlled trial (RCT) design) and in this study, researchers will attempt to replicate those results across a broader set of school sites. The team will also investigate how differences across students, teachers, and sites explain variability in the program's observed effects, and examine the successes and challenges in replicating and reproducing the program's processes and impacts across the diverse sample of teachers and sites.
Project Activities: The research team will use a RCT to examine the impact of DLL on literacy achievement for two first-grade cohorts of Spanish-speaking ELs who are struggling readers. Researchers will randomly assign students to experimental conditions across 30 schools, and participate in pre- and post-tests of their Spanish and English reading skills using intervention-specific and standardized, norm-referenced measures. The team will collect measures of program implementation, and will examine factors hypothesized to moderate and mediate outcomes. In addition, researchers will compare the second grade literacy achievement of DLL students to a random sample of ELs who did not receive services, to look for any ongoing or later-emerging effects after the intervention is completed.
Products: Researchers will l provide evidence about the efficacy of the DLL reading intervention on improving literacy outcomes and closing achievement gaps for young Spanish-speaking ELs who are struggling readers. The research team will also produce peer-reviewed publications, conference presentations, and engage in web and social media outreach to disseminate their findings.
Setting: This study will include 30 schools across four states (California, Wisconsin, Illinois, and Texas).
Sample: The school sample includes 30 schools whose collective demographic profiles are roughly similar to that of the full population of 82 schools where DLL is implemented. Within schools, teachers are selected if they have already been trained to administer DLL, teach at least four students per day, keep complete records on each student as a basis for instruction, use the Spanish version of the DLL Observation Survey (OS), communicate with parents and teachers, submit data to the teacher leader, and monitor the progress of students who have exited the program. The student sample will include the lowest performing ELs in each of two first-grade cohorts, as determined by a series of literacy assessments administered as students begin the school year. Each cohort will contain at least 360 first grade students (12 per school), for a total of 720 students.
Intervention: DLL provides one-on-one tutoring in Spanish for first-grade Spanish-speaking students who are receiving literacy instruction in Spanish and experiencing substantial challenges with reading and writing. The program offers daily 30-minute instruction for students performing in the bottom 20 to 25 percent of their class. Trained literacy teachers provide intensive, individualized support on letter and sound relationships, spelling patterns, comprehension, producing oral language, and increasing vocabularies. DLL lessons are administered as a complement to students' regular-classroom bilingual literacy instruction, with the goal of raising struggling readers' literacy skills to the level of their peers. Each student progresses through DLL at an individualized pace, completing the program 12–20 weeks after initial assignment. The DLL program also includes rigorous teacher training, instructional standards, and ongoing professional development. Participating classroom teachers receive one-on-one support from a teacher leader, a minimum of four professional development sessions per annum, and multiple classroom observations per year from both teacher leaders and colleagues.
Research Design and Methods: Researchers will employ a multisite, student-level randomized research design in 30 schools across four states. The team will implement the randomized control trial (RCT) over the first two project years. The researchers will follow two cohorts of students who meet the selection criteria and are enrolled in first grade in Fall 2016–17 or Fall 2017–18. In each cohort, all selected students receive the intervention, but they are randomly assigned to a first group or a delayed group, with the latter serving as a control condition for the former. Each participating student will progress through DLL at an individualized pace, completing the program 12–20 weeks after initial assignment. As students enter and exit the DLL treatment condition, they will complete three pre- and post-test standardized reading measures (two in Spanish, one in English).
Researchers will also conduct site visits and classroom observations to gather information on implementation fidelity in classrooms and schools. To evaluate the longer-term impacts of DLL on reading achievement, the team will continue to collect outcome measures from students through the end of the 2018–19 school year (third- and second-grade for the first and second cohorts, respectively). Researchers will dedicate the final two years of the project (June 2018 through June 2020) to data analyses and dissemination.
Control Condition: This study uses a delayed treatment design where all students receive the intervention, but students in the delayed group serve as controls for the initial treatment group and only rotate into the treatment condition once the first wave of students completes it.
Key Measures: The researchers will use three outcome measures to determine the impact of DLL on literacy achievement: the Iowa Test of Basic Skills (ITBS), the Spanish version of the ITBS (Logramos), and an intervention-specific Spanish-language literacy assessment called the Instrumento de Observación (IdO). The first two assessments are standardized, norm-referenced, and independent from the intervention, while the third is specific to the intervention and is supported by validity research and evidence from prior studies. Each instrument has five to seven subtests of specific reading skills or practices. In addition to student reading outcomes, data from surveys, activity logs, observations, and interviews will be used to quantify implementation fidelity across sites. Researchers will use pre-established thresholds of low, medium, and high fidelity in analyses linking level of implementation with student performance. Researchers will also explore moderating effects from school-, teacher-, and student-level contextual variables such as geographic region, teacher experience, and initial student literacy levels.
Data Analytic Strategy: Researchers will measure the impact of DLL on literacy outcomes and achievement gaps using multilevel regression models with students nested within sites. The team will add student-, teacher-, and school-level covariates to this baseline model to explore moderation effects on the program's impact. The researchers will also use a latent variable mixture modeling approach to estimate the effects of non-compliance on final effect estimates. They will use descriptive analyses to summarize information about fidelity of implementation, and regression models to explore potential mediation effects from both teachers and students.