DESCRIPTION (Adapted from the abstract): The proposed research will build on previous and current studies of computer-based remediation for reading disabilities wherein 2nd - 6th grade children read interesting stories on the computer for a half hour each day during the irremedial-reading or language-arts class time. Children are trained to target difficult words in the stories with a mouse for immediate orthographic and speech feedback. The computer emphasizes important relations between word or sub-word letter patterns and their corresponding speech sounds by displaying the patterns in reverse video while a high-quality speech synthesizer simultaneously pronounces them. Compared to control groups, trained subjects' average gains across a semester were about four times greater in phonological decoding of nonwords and about two times greater in word recognition. However, individual-differences analyses of disabled readers' gains revealed interactions between deficit-severity, initial level of phonological awareness, and assigned level of feedback segmentation (whole-word, syllable, or onset-rime units). One goal of the proposed research is to confirm these pedagogically important interactions in a year-long training study with more stable measures of performance. The major goal of the research is to test hypothesized additional benefits of computer-based training in phonological awareness, because the previous subjects' gains in phonological decoding and word recognition were strongly related to their initial level of phonological awareness at pre-test. Training effects will be compared between subjects who use only computer-based reading with speech feedback, and subjects who spend the first 4-6 weeks on computer-based phonological-awareness training before reading, followed by reading on the computer with concurrent phonological-awareness exercises. Total training time and experimenter contact will be the same for both groups so that any unique benefits of phonological- awareness training will not be confounded by extraneous variables. Main effects and interactions will be assessed for phonological- awareness vs. only reading on the computer and the three different levels of feedback segmentation in reading. Individual differences in subjects' gains for phonological awareness, phonological decoding, word recognition, spelling, and reading comprehension will be assessed at the end, and one or two years after training through the use of hierarchical linear models.