Identifying individual children who meet criteria for learning disability (LD) has plagued research and practice since the origin of the concept of LD. In this project, we leverage the special statistical and clinical expertise of our team and advances in statistical computing and analytic models, simulation, and meta-analysis to continue and extend a long history of research on the classification and definition of LD. We evaluate the reliability of different approaches to identification; the validity of classifications based on intervention response; and the integration of research on classification, executive functions, and intervention.
In our recent research, we have shown the following:
- Different methods of classifying children as being at risk for reading failure, including slopes and intercepts from response to intervention (RTI) models, benchmarks, and normative cutpoints, show low agreement.
- IQ is not a good predictor of RTI. A meta-analysis of using IQ to predict RTI found that this model accounted for only about 1% of the variance.
- Several current methods for identifying specific LD from profiles of cognitive strengths and weakness lack sensitivity and, if generally applied, would identify many nonimpaired children as being impaired or would identify children with an incorrect impairment. We demonstrated this finding by using simulation studies.
- The reported effectiveness of comprehension interventions for adolescents is inconsistent. A meta-analysis of the effectiveness of comprehension interventions for adolescents found large effects for experimenter-developed measures but only moderate effects for standardized, normed measures.
In the current funding cycle, the overarching themes of Project 1 remain identification and integration. This project will use statistical simulation and real data to research the classification and identification of children with LD and extend this work to English learners (ELs). This project will also build on previous syntheses and meta-analyses to examine ELs and students with comorbid disabilities. This initiative integrates research across projects, approaches, frameworks, settings, and populations.
- Investigate (a) whether LD is inherently dimensional or nondimensional by developing and evaluating new approaches to identification that overcome known problems with existing approaches and (b) the application of these methods with a large sample of ELs to address fundamental questions about how to identify LD in ELs with persistent academic difficulties
- Conduct research syntheses and meta-analyses examining (a) the history of LD identification, (b) predictors of intervention response in comorbid disabilities, and (c) cognitive similarities and differences between comorbid and singular disabilities
View publications related to this project in our research library.