Purpose: The purpose of this project is to use an integrative-data-analysis approach with seven prior randomized controlled trial (RCT) studies to (a) examine the degree to which interventions designed to promote children's reading-related and language skills also positively impact children's self-regulation, (b) identify child and intervention factors that contextualize this effect, and (c) examine the degree to which effects on self-regulation are mediated by children's gains on the target skill of the interventions to better identify mechanisms responsible for gains in self-regulation.

Project Activities: The researchers will use integrative data analysis (IDA) to combine self-regulation outcomes from seven prior randomized studies originally conducted to determine the impacts of specific instructional activities on the reading-related outcomes of children at risk for reading difficulties. IDA requires reconfiguration and recoding of data to be able to develop a common outcome across studies. In this project, this will include measures of self-regulation, fidelity of implementation of the intervention, children's participation in the intervention, and child outcomes on standardized measures of reading-related skills. These data will be used to determine if these academically focused interventions also have an impact of children's self-regulation and, if so, for whom and under what conditions.

Products: The products resulting from this project will include information about the presence, size, and scope of impacts of language and reading-related interventions on children's self-regulation outcomes, including the degree to which characteristics of children (grade, sex, race, ethnicity, initial levels of self-regulation) and characteristics of the interventions (specific focus, fidelity, received dose) enhance or limit those impacts. The project will provide important information about the causal relations between academic skills and self-regulation, which will help inform the future use and development of education practices. Findings will be shared with educators and policymakers through multiple channels, and findings will be shared with researchers through peer-reviewed publications and presentations.

Structured Abstract

Setting: Children included in the seven prior randomized studies to be used for this project were recruited primarily from preschools and public schools characterized by higher poverty levels (more than 50 percent of students in school eligible for free/reduced-price lunch). Most of the studies were conducted in Florida, but several of the studies were also conducted with schools from California, Texas, Kansas, and New Mexico. Locations included urban and rural areas.

Sample This project will use data from seven prior randomized studies that had as their primary focus examination of the efficacy of academic interventions on the language, reading, and reading-related skills of children in preschool through fourth grade. These original studies focused on children at risk for reading-related difficulties. Across the seven studies, 9,964 children participated, representing diverse racial and ethnic groups.

Intervention: The core features of the instructional approaches in the seven studies included either exclusively or significant amounts of small-group delivery of explicit and relatively scripted instruction. Instructional targets (vocabulary, syntax, phonological awareness, print knowledge) were taught across a variety of contexts that included systematic practice, corrective feedback, and scaffolding for mastery learning. Most instruction involved modular units of 20- to 30-minute, daily lessons that collectively lasted from nine weeks to the entire school year.

Research Design and Methods Each of the seven studies used a randomized design, often with some level of matching prior to randomization. Two studies were cluster-randomized studies of curriculum (randomization at school level), and the remaining five studies involved individual randomization of children who met specific qualification criteria. All studies included assessments of participation and measurement of fidelity of implementation.

Control Condition Across the seven studies, the control condition was standard classroom practice or, in the case of curricular manipulations, the existing curriculum and professional development in use in a preschool or school (business-as-usual control groups).

Key Measures Across studies, the primary outcome for this project includes various teacher reports of children's self-regulation, including reports of behaviors associated with disruptive behavior disorders and social competence. Four of the studies include performance-based assessments of children executive function (inhibitory control, working memory).

Data Analytic Strategy The research team will use an integrative data analysis (IDA) approach, combining data across studies using moderated nonlinear factor analysis (MNLFA). MNLFA allows the simultaneous evaluation and correction of potential sources of bias in outcome measures (different measurement responses due to study or characteristics of the sample), and it allows the simultaneous evaluation of continuous and categorical variables. The output of MNLFA is a measurement model in which the outcome has the same meaning across studies and across populations within and across studies. The a priori measurement model for this project involves a bifactor (S-1) approach for teacher-reports of self-regulation in which hyperactive/impulsive behaviors serve as the General (or reference) factor, and performance-based measures serve as an additional set of indices for the General factor. This measurement model will allow specification of both unique and common aspects of report- and performance-based measures of self-regulation. The primary hypotheses concern the impact of the interventions on self-regulation immediately following the intervention and, when available, at follow-up. Analysis of moderation and mediation will examine the contextual limits and mechanisms for the effects on self-regulation. IDA retains the internal validity of the original studies because intervention contrasts are conducted within study; however, the combination of data across studies substantially increases statistical power. All analyses will involve multi-level models that consider the nested structure of the data (i.e., children nested within classrooms and schools). Mediation will be tested using a bootstrapping procedure within the multi-level framework.

Cost Analysis: The research team will determine costs for each intervention, including personnel effort, professional development, and materials for schools (local, state-wide). They will compute cost-effectiveness as the ratio of costs per student for each standard deviationunit of effect size for each intervention.

Related IES Projects: Examining Effective Intervention Targets, Longitudinal Intensity, and Scaling Factors for Pre-K to 5th Grade Student Comprehension (R305F100027); Generating Large and Sustained Impacts on Early Language Skills: Evaluation of Timing and Duration of Intervention (R305A160241)