The proposed reading comprehension training software, 3D-Readers, will fill the current need in schools for software the remediates reading problems beyond the decoding level. It is based on the hypothesis that poor comprehenders benefit from training in, and practice on, two types of metacognitive strategies. The proposed, interactive software will use highly visualizable third through tenth grade science texts to: 1) Diagnose poor reading comprehenders. 2) Train them in the verbal strategy of generating questions, and then score generated questions and constructed answers with a flexible essay-scoring algorithm. 3) Train them in creating visual mental models. 4) Create individualized cognitive profiles using a powerful neural network algorithm. A computer-generated analysis of each reader's unique cognitive profile will ascertain which variables are most responsive to change. This profile analysis will aid teachers by suggesting remediations most likely to enhance students reading comprehension skills. A pilot study with an experimental and treated control group will establish the feasibility embedding the metacognitive strategies in computerized text. The long term goal is to offer affordable, interactive software to at-risk readers that will train them in generative reading strategies they will use throughout their lives.
3D-Readers is reading comprehension training and assessment software that will train metacognitive strategies on 4th through 10 grade level texts. The software's highly visualizable science texts will be appropriate for use in public schools, home schools, learning centers, and even in some traumatic brain injury remediation situations. The innovative use of neural network technology and essay grading algorithms make the system highly flexible yet accurate.
