This study aims to show that training using a computer game incorporating an audio-visual phoneme discrimination task with phonological units, presented simultaneously with orthographic units, might improve literacy skills. Two experiments were conducted, one in secondary schools with dyslexic children (Experiment 1) and the other in a speech-therapy clinic with individual case studies (Experiment 2). A classical pre-test, training, post-test design was used. The main findings indicated an improvement in reading scores after short intensive training (10 h) in Experiment 1 and progress in the reading and spelling scores obtained by the dyslexic children (training for 8 h) in Experiment 2. These results are discussed within the frameworks of both the speech-specific deficit theory of dyslexia and the connectionist models of reading development.

Two experiments were conducted to test the hypothesis that sentence processing is an essential mediatory skill between word recognition and text comprehension in reading. In Experiment 1, a semantic similarity judgement task was used with children from Grade 2 to Grade 9. They had to say whether two written sentences had the same (or very similar) meanings or whether the meanings of the two sentences were very different. As expected, performance improved with age both on the high-frequency words and with increasingly complex sentences. In Experiment 2 with children in Grade 3, scores in written sentence comprehension and vocabulary made the most important unique contribution to the reading comprehension of an expository text. The results are discussed first, in the light of a general framework in which sentence-level skill is at the core of reading comprehension and second, with reference to the issue of reading assessment.The simple view of reading (Gough & Tunmer, 1986;Hoover & Gough, 1990) defines reading comprehension as a product of word-level decoding and linguistic comprehension (an oral language construct). However, text comprehension in reading implies multicomponent skills, from basic skills involving low-level word reading processes such as phonological awareness, decoding and written word recognition through to higherorder comprehension processes that relate to the construction of coherence and specifically involve the making of inferences and, more generally, background knowledge about the texts to be read Rapp, van den Broeck, McMaster, Kendeou & Espin, 2007). In a simple bottom-up view of reading (i.e. from word level to text level), which finally leads to the construction of the mental model which represents the text, the reader has to read a sequence of sentences and construct the meaning of each of them locally.The general aim of this study is to investigate the importance of sentence-level skill as an intermediate and essential skill situated between word recognition and text comprehension. Sentence-level skill in reading is therefore mobilised and assessed in two experiments. First,

No abstract

The goal of the experiment was to examine the relevance of a new French web-based assessment, Tinfolec (Test INFOrmatisé d’évaluation de la LECture), the aim of which is to evaluate the reading abilities of children in primary grades. The participants were 1,016 children from Grades 2 to 5. They completed the five tasks of Tinfolec designed to assess the efficiency of the two procedures used to identify written words (the nonlexical route and the lexical orthographic route). We tested the reliability and validity of the new tool in a subsection of this sample. Correlational analyses provided evidence of the reliability and validity of Tinfolec. The results are consistent with the conventionally observed effect of lexical factors (length, consistency, and frequency) on written word processing. The results confirm the relevance of the proposed tasks. The study produced promising results and would allow practitioners to perform online assessments of reading skills.