Response to fluency-oriented intervention of Dutch poor readers

Dysfluency is the major characteristic of dyslexia in languages with a relatively shallow orthography, particularly when unfamiliar words have to be processed. The present study investigated to what extent slow identification of unfamiliar words can be speeded up and whether the expected variation in response-to-intervention can be related to differences in the phonological recoding skills in order to gain insight in the core deficits of dyslexia. The intervention involved four months of continued individual practice of fluency after maximal accuracy was achieved. Response-to-intervention was made operational in terms of individual learning curves, derived from multilevel analysis with repeated measures of response times, number of trials and accuracy as dependent variables. Increasing fluency was indicated by response times getting shorter. An unselected group of poor grade 2 readers who received the intervention in addition to regular classroom practice improved their reading compared to a matched control group who received no additional training. As expected, there were substantial individual differences between students in terms of response-to-intervention (RTI). Good and poor treatment responders progressed comparably when the training was done in the relatively easy training condition with closed word sets of four letter words with one consonant cluster, repeating the overlapping cluster + vowel. Increasing the difficulty level of the training condition to word sets overlapping in consonant cluster only, resulted in substantial individual differences in all three dependent variables. In fact, the poor responders did not improve any more whereas the poor responders continued to do so. This particular characteristic of each student's individual learning curve appeared to be highly predictive of the overall response-to-intervention or intervention outcome. The contribution of this finding to the understanding of dyslexia is discussed in terms of RTI.