Predicting response to early reading intervention for children at risk for dyslexia - the importance of pronunciation accuracy.

This study aimed to examine the extent to which pretest measures predicted response to an early reading intervention for children at risk for dyslexia. Predictors were studied in both first-language (L1) speakers and second-language (L2) learners. The pretest measures assessed various reading-related abilities, including pronunciation accuracy. Data from the experimental group in a trial of a reading intervention of 76 sessions for first-grade students at risk for dyslexia were analyzed. Outcomes were examined for 101 participants, including 68 L1 speakers and 33 L2 learners of Danish, based on their posttest performance on reading of words and non-words. Pretest measures of decoding, letter knowledge, letter-sound knowledge, phonological awareness, rapid automatized naming, vocabulary, and pronunciation accuracy were evaluated as predictors of reading outcomes. The strongest predictors of children's reading outcomes were phonological awareness, letter knowledge, letter-sound knowledge, and pronunciation accuracy. A relatively good classification of children with subsequent poor or adequate response to the intervention could be obtained based on pretest measures of letter-sound knowledge and pronunciation accuracy. Pronunciation accuracy added uniquely to the prediction of overall variance in reading outcomes for L1 speakers and L2 learners and to the correct early classification of children as poor or adequate responders. The results suggest that an acceptable level of prediction of at-risk children's response to early reading intervention can be obtained by means of a few pre-intervention measures. This study provides evidence for the added predictive value of a measure of pronunciation accuracy for L1 speakers and L2 learners.