Predictors of Speech Severity and Inconsistency Over Time in Children With Token-to-Token Inconsistency.

PURPOSE: The aim of this study was to describe and explain changes in severity of speech sound disorder (SSD) and token-to-token inconsistency in children with high levels of inconsistency. METHOD: Thirty-nine children (aged 4;6-7;11 [years;months]) with SSDs and high levels of token-to-token inconsistency were assessed every 6 months for 2 years (i.e., five assessment points). Growth modeling was used to assess relations among therapy support, receptive vocabulary, severity, and inconsistency over time. RESULTS: Children with the most severe SSDs and highest levels of token-to-token inconsistency showed the smallest improvements in speech accuracy over time. Therapy support did not predict changes in speech accuracy or token-to-token inconsistency over time. Receptive vocabulary (measured at the outset of the study) was also a significant predictor of speech accuracy and inconsistency. CONCLUSIONS: These findings suggest that an immediate start to intervention (rather than a wait-and-see approach) is recommended for children with inconsistent speech error patterns. The results also highlight the value of developing vocabulary knowledge in addition to improving speech accuracy for some children with inconsistent speech production.

Predicting Dyslexia in Children With Developmental Language Disorder.

Purpose In this study, we examine how well kindergarten letter identification and phonological awareness predict 2nd grade word reading and dyslexia in children with developmental language disorder (DLD) and their age- and grade-matched peers with typical language (TL). Method We employ (a) logistic regression to determine how letter identification and phonological awareness predict dyslexia, that is, dichotomous categorization of good or poor word reading, in children with DLD and TL and (b) quantile regression to determine how letter identification and phonological awareness are associated with word reading abilities on a continuum in these groups of children. Results Logistic regression revealed that letter identification was the only significant, unique kindergarten predictor of dyslexia in 2nd grade children with DLD, when compared to phonological awareness. In children with TL, both kindergarten letter identification and phonological awareness significantly predicted dyslexia in 2nd grade. Quantile regression revealed that kindergarten letter identification was a stronger predictor of 2nd grade word reading for average and lower achieving word readers with DLD and their peers with TL compared to higher performing readers. Phonological awareness was weakly associated with word reading across the full continuum of word reading abilities in children with DLD. Conclusion Letter identification is a more accurate predictor of poor word reading and dyslexia than phonological awareness in kindergarten children with DLD, which has important implications for recent U.S. legislation around early identification of dyslexia in all children.