We review significant empirical studies of the arithmetic abilities of children with dyslexia. These studies suggest that the academic impairments of children with dyslexia are not limited to reading and spelling, but also include aspects of mathematics. A consistent finding across a number of studies is that children with dyslexia have difficulty recalling number facts. The results of the reviewed studies are analysed, both in terms of the weak phonological representations hypothesis, and the triple-code theory of mathematical cognition. It is suggested that the phonological processing deficits of individuals with dyslexia impair aspects of mathematics that rely on the manipulation of verbal codes (e.g. counting speed, number fact recall), whilst other aspects of mathematics that are less reliant on verbal codes (e.g. estimation, subitising) are unimpaired. Suggestions for testing this hypothesis are put forward.

No abstract

This study examines the relationships between phonological awareness, visualspatial sketchpad (VSSP) functioning and arithmetic attainment in young children. A sample of 42 children had their VSSP functioning and phonological awareness assessed when they were 5 years old. Approximately 12 months later their nonverbal reasoning, vocabulary, arithmetic, and reading attainment were assessed. Together, VSSP functioning, phonological awareness, vocabulary, and nonverbal reasoning predicted 41% of the variation in the children's arithmetic attainment. Only phonological awareness and VSSP functioning were significant independent predictors. In contrast, only phonological awareness was a significant independent predictor of reading attainment. These findings are consistent with phonological awareness influencing both the development of reading and arithmetic, whilst VSSP functioning only impacts on arithmetic development.When children solve arithmetic problems they must generate and store either nonverbal or verbal representations of the quantities. Huttenlocher, Jordan, and Levine (1994) proposed that very young children use nonverbal mental models. Rasmussen and Bisanz (2005) further developed this proposal, arguing that young children's mental representations of quantities consist of a visual array of tokens that can be manipulated to perform arithmetic operations. They hypothesised that these visual representations are stored in the visual-spatial sketchpad (VSSP). Rasmussen and Bisanz suggest that as children progress through school they increasingly use verbal representations of quantities (i.e., number words), which are phonologically coded and can Correspondence should be addressed to Fiona Simmons,

The reading comprehension abilities of a group of dyslexic university students were compared with those of non-dyslexic university students. A 655-word passage, followed by literal and inferential questions, was used to measure reading comprehension. The text was designed to be syntactically complex, yet place relatively modest demands on decoding skills. Although dyslexic students performed at a similar level to the non-dyslexic students on the literal questions, their performance on the inferential questions was poorer. An index of the participants' ability to make inferences was calculated by subtracting the inferential question score from the literal question score. The groups differed significantly on this measure, indicating that the dyslexic students were specifically impaired in constructing inferences when processing complex text. It was concluded that dyslexic students in higher education have reading comprehension difficulties that cannot be accounted for by an inability to decode individual words in the text. The possible contribution that poor lexical automaticity and an impaired working memory make to this impairment is discussed. The implications for the assessment and support of dyslexic students are considered.

Identifying dyslexia in adulthood presents particular challenges because of complicating factors such as acquisition of compensatory strategies, differing degrees of intervention and the problem of distinguishing dyslexic adults from those whose literacy difficulties have non-cognitive causes. One of the implications is that conventional literacy measures, per se, do not provide a satisfactory basis for screening for dyslexia in adulthood as some dyslexic adults have above-average literacy skills and some non-dyslexic adults have very poor literacy skills. This study examined an alternative approach to dyslexia screening, using three tests that depend heavily on phonological processing, lexical access and working memory, but which are not conventional measures of literacy. Using these tests, which are computer delivered, 70 dyslexic adults from three different types of educational institution were compared with 69 non-dyslexic adults from the same institutions. The results showed that the dyslexic and non-dyslexic groups were significantly different on all three computer-based tests, with an average effect size of 1.55. Adaptive versions of these tests were then created to reduce overall administration time for the suite to about 15 minutes. Analysis showed that the combined scaled scores from the adaptive versions of the three tests significantly discriminated the dyslexic from the non-dyslexic group with an increased effect size of 2.07 and with a sensitivity rate of 90.6% and a specificity rate of 90.0%. It was concluded that this approach is a valid and useful method of identifying dyslexia in adulthood, which, given the ease of administration to large numbers of adults, has noted advantages for education and employment.

The abilities of 19 adult students with dyslexia and 19 students without dyslexia to recall number facts were compared. Despite being matched for estimated IQ, the dyslexic students were less accurate than the non-dyslexic students when answering subtraction and multiplication questions. When the dyslexic students answered addition and subtraction questions correctly they did so more slowly than the non-dyslexic students. It was concluded that dyslexic adults have greater difficulty recalling number facts than non-dyslexic adults. No statistically significant correlations were found between the dyslexic students' performance on the tests of number fact recall and their performance on tests of auditory-verbal and visual-spatial memory.

Gender differences in the relationships between working memory (short-term storage and combined storage and processing) in both the visuo-spatial and verbal domains and children's alphabet transcription and text writing abilities were investigated. Data from 81 children (43 males) aged between 5;2 to 8;5 revealed no significant group differences between boys and girls in working memory or writing performance. However, individual differences analyses demonstrated variation associated with age and gender in the memory skills underpinning writing. Regression analyses revealed that verbal short-term memory abilities predicted the alphabet transcription skills of boys but not girls. Although visuo-spatial short-term memory predicted writing quality in both genders, predictors of writing fluency differed with verbal working memory skills predicting boys' writing fluency and visuo-spatial short-term memory predicting writing fluency in girls. The need to consider gender differences more critically from the perspective of individual differences in cognitive skills underpinning writing development and the strategic application of these skills during writing is discussed.

The extent that phonological, visual-spatial STM and non-symbolic quantitative skills support the development of counting and calculation skills was examined in this 14-month longitudinal study of 125 children. Initial assessments were made when the children were 4:8. Phonological awareness, visual-spatial STM and non-symbolic approximate discrimination predicted growth in early calculation skills. These results suggest that both the approximate number system and domain-general phonological and visual-spatial skills support early calculation. In contrast, only performance on a small non-symbolic quantity discrimination task (where the presented quantities were always within the subitising range) predicted growth in cardinal counting skills.These results suggest that the development of counting and calculation are supported by different cognitive abilities.