Edited by:
Reviewed by:
*Correspondence:
This article was submitted to Educational Psychology, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Reading and spelling performance have a significant correlation with number transcoding, which is the ability to establish a relationship between the verbal and Arabic representations of numbers, when a conversion of numerical symbols from one notation to the other is necessary. The aim of the present study is to reveal shared and non-shared mechanisms involved in reading and writing of words and Arabic numerals in Brazilian school-aged children. One hundred and seventy-two children from second to fourth grades were evaluated. All of them had normal intelligence. We conducted a series of hierarchical regression models using scores on word spelling and reading single words and Arabic numerals, as dependent variables. As predictor variables we investigated intelligence, the phonological and visuospatial components of working memory (WM) and phonemic awareness. All of the writing and reading tasks (single word spelling and reading as well as number reading and number writing) were significantly correlated to each other. In the regression models, phonological WM was specifically associated to word reading. Phonemic awareness was the only cognitive variable that systematically predicted all of the school skills investigated, both numerical and word tasks. This suggests that phonemic awareness is a modular cognitive ability shared by several school tasks and might be an important factor associated to the comorbidity between dyslexia and dyscalculia.
Reading and math performance in school-aged children are related in important ways. In this study, we set out to investigate shared and non-shared mechanisms involved in word reading/spelling and number transcoding abilities. Genetically informed studies indicate that performance in standardized math, spelling, and reading achievement tests substantially correlate at the phenotypic and genetic levels, both in typically and atypically developing children (
Other evidence points to dissociations between reading and math performance. Pure cases of multifactorial dyslexia and dyscalculia hint at non-shared specific cognitive mechanisms that may characterize distinct entities or subtypes (
Shared and non-shared mechanisms can be identified at the neural level (
In a systematic review,
Which are the possible shared and non-shared mechanisms at the cognitive level of description? In order to search for an answer to this question, it is important to investigate outcome measures shared by reading and arithmetics. Standardized achievement tests are not good indicators because they evaluate abilities at different complexity levels, recruiting semantic and reasoning processes in different degrees. We focused then on some relatively low-level abilities of word reading/spelling and basic numerical processing (Arabic number reading and Arabic number writing).
These basic, decontextualized and less semantically loaded abilities have been consistently implicated in reading and math performance both in typical and atypical populations. In the case of reading, for example, it has been assumed that reading at the word level is an important precursor of more advanced mechanisms of reading comprehension (
Focusing on the basic abilities at the lexical and numerical processing levels, we also hope to uncover meaningful patterns of association and dissociation between cognitive variables and reading/spelling- and math-related performances in early school age. According to
The picture is, however, complicated by the fact that arithmetics-related abilities are more heterogeneous than reading-related abilities. Research increasingly shows that some aspects of arithmetics and number processing may be dependent on verbal processes, sharing mechanisms with reading learning and dyslexia (
Following
Intelligence is an important long-term predictor of school achievement assessed with standardized omnibus tests (
The role of intelligence in number reading and writing has been the focus of less research. However, extant data indicate that general cognitive ability is significantly associated with Arabic number reading and dictation in the school-age population (
Regarding visuospatial WM,
The association between visuospatial WM and reading/spelling is unclear and more studies are required. A recent investigation explored the complete WM profile of children with poor reading ability (
Phonological WM is the mechanism involved in the temporary storage and manipulation of verbal items (
Phonological awareness can be investigated by means of tasks that demand the distinction between the sounds that constitute words, such as rhyme detection and blending isolated sounds to create words (
A question that demands further discussion lies in the fact that the phonological complexity of the stimulus might be a confounding factor to mask the association to reading (
Children with developmental dyslexia, who perform poorly on phonological processing tasks, such as phonemic awareness, frequently exhibit deficits in mathematics. According to the weak phonological representation hypothesis (
The aim of the present study is to investigate shared and non-shared mechanisms involved in reading and writing words and Arabic numerals in school-aged children. Our main goal was to disentangle the role of phonemic awareness and its impact on lexical and numerical tasks controlling for cognitive variables which may have an important impact. We hypothesize that even after controlling for the influence of intelligence and broader reading and writing skills, phonemic awareness would be an important predictor since all of these tasks involve some sort of verbal processing. In order to investigate that, we conducted a series of hierarchical regression models using scores of reading and writing of single words and Arabic numbers tasks, as dependent variables. As predictor variables we investigated intelligence, the phonological and visuospatial components of WM and phonemic awareness skills.
The study was approved by the local research ethics committee (COEP–UFMG). Children participated only after informed consent was obtained. It was obtained in written form from parents and orally from children.
We have assessed 207 children from second to fourth grades of public schools from Belo Horizonte, Brazil. Data collection took place in the participants’ schools. We excluded one child who did not complete the entire battery, four children due to low intelligence (performance on Raven’s Colored Progressive Matrices below one standard deviation from the mean), and 30 children were excluded from further analysis, because either they had a poor
At first, the intelligence (Raven’s CPM), word spelling (Brazilian School Achievement Test – TDE) and number transcoding (Number writing task) were evaluated in small groups of approximately nine children. Subsequently, we tested number transcoding (Number reading task), word reading (reading subtest of Brazilian School Achievement Test – TDE), phonemic awareness (Phoneme Elision), and WM (Corsi Blocks and Digit Span).
General intelligence was assessed with the age-appropriate Brazilian validated version of Raven’s Colored Matrices (
The Teste de Desempenho Escolar (TDE,
To evaluate number transcoding, children were instructed to write the Arabic forms of dictated numbers. This task consists of 40 items, up to four digits (3 one-digit numbers, 9 two-digit numbers, 10 three-digit numbers, and 18 four-digit numbers). The one- and two-digit numbers were classified as “lexical items” (12 items), and the other 28 items require the use of algorithm-based rules in order to be written (
A total of 28 Arabic numbers with one to four digits were printed in a booklet and presented to children one at a time. Children were instructed to read them aloud. Items were grouped into three categories according to their complexity, indexed by the number of transcoding rules established by the ADAPT model. The three- and four-digit numbers were chosen to avoid presenting numbers with very strong lexical entries and to maintain the focus on syntactic complexity. This task has been used in previous studies, and the consistency of this task was
This is a widely accepted measure of phonemic awareness (
This test is a measure of the visuospatial component of WM. It is constituted by a set of nine blocks which are tapped, in a certain sequence, by the examiner. The test starts with sequences of two blocks and can reach a maximum of nine blocks. We used the forward and backward orders according to
Verbal short-term memory was assessed with the Brazilian WISC-III Digits subtest (
Raw scores were
To explore the general pattern of association between the cognitive variables and reading and writing of numbers and words, we investigated the correlations between them (
Correlations between the neuropsychological measures.
Word spelling | Word reading | Number writing | Number reading | Phoneme elision | Digit span | Corsi blocks | |
---|---|---|---|---|---|---|---|
Raven | 0.449∗∗ | 0.313∗∗ | 0.401∗∗ | 0.411∗∗ | 0.371∗∗ | 0.359∗∗ | 0.329∗∗ |
Word spelling | 1 | 0.719∗∗ | 0.593∗∗ | 0.562∗∗ | 0.557∗∗ | 0.317∗∗ | 0.155∗ |
Word reading | 1 | 0.503∗∗ | 0.598∗∗ | 0.693∗∗ | 0.395∗∗ | 0.149 | |
Number writing | 1 | 0.762∗∗ | 0.483∗∗ | 0.271∗∗ | 0.244∗∗ | ||
Number reading | 1 | 0.602∗∗ | 0.287∗∗ | 0.238∗∗ | |||
Phoneme elision | 1 | 0.366∗∗ | 0.253∗∗ | ||||
Digit span | 1 | 0.354∗∗ | |||||
Corsi blocks | 1 |
As can be seen in
To further explore the association between these variables and to have a more fined-grained perspective regarding the predictive power of them on each of the verbal and numerical skills, we conducted separate hierarchical regression models.
We calculated regression models including intelligence and the analogous variables in the first step and the other cognitive variables in the second one. By doing so, we aimed at investigating the pattern of association between these variables, once we had controlled for more general cognitive skills. Variance for each kind of number task (e.g., reading or writing) was predicted by the homologous tasks with words (e.g., reading or writing) and vice-versa (
Regression analysis for number reading (adjusted
Predictor | Beta | Partial |
Significance | |
---|---|---|---|---|
Intercept | –2.387 | 0.018 | ||
Raven | 0.196 | 3.194 | 0.002 | |
Word reading | 0.327 | 4.142 | <0.001 | 0.413 |
Digit span | –0.03 | –0.469 | 0.64 | Excluded |
Corsi blocks | 0.055 | 0.904 | 0.367 | Excluded |
Phoneme elision | 0.303 | 3.747 | <0.001 | 0.045 |
Regression analysis for word reading (adjusted
Predictor | Beta | Partial |
Significance | |
---|---|---|---|---|
Intercept | 0.441 | 0.660 | ||
Raven | –0.034 | –0.575 | 0.566 | |
Number reading | 0.279 | 4.143 | <0.001 | 0.363 |
Digit span | 0.151 | 2.614 | 0.010 | 0.018 |
Corsi blocks | –0.099 | –1.746 | 0.083 | Excluded |
Phoneme elision | 0.482 | 7.136 | <0.001 | 0.168 |
Regression analysis for number writing (adjusted
Predictor | Beta | Partial |
Significance | |
---|---|---|---|---|
Intercept | –5.357 | <0.001 | ||
Raven | 0.138 | 2.042 | 0.043 | |
Word spelling | 0.421 | 5.557 | <0.001 | 0.374 |
Digit span | 0.019 | 0.291 | 0.804 | Excluded |
Corsi blocks | 0.096 | 1.509 | 0.867 | Excluded |
Phoneme elision | 0.197 | 2.71 | 0.007 | 0.026 |
Regression analysis for word spelling (adjusted
Predictor | Beta | Partial |
Significance | |
---|---|---|---|---|
Intercept | 4.421 | <0.001 | ||
Raven | 0.187 | 2.984 | 0.003 | |
Number writing | 0.369 | 5.557 | <0.001 | 0.405 |
Digit span | 0.046 | 0.737 | 0.462 | Excluded |
Corsi blocks | –0.087 | –1.454 | 0.148 | Excluded |
Phoneme elision | 0.309 | 4.721 | <0.001 | 0.070 |
In the present study, we investigated the cognitive variables that underlie the performance of reading and writing skills for both numbers and words in a sample of Brazilian school-aged children. Our results can be summarized into two mains topics: the specific influence of phonemic awareness on reading and writing words and numbers and the impact of non-verbal intelligence on them. We have found a prominent role of phonemic awareness which was consistently associated to all the reading and writing modalities we have assessed. As far as we know, this is the first study to simultaneously investigate these four skills and the cognitive variables related to them.
Phonemic awareness is an important underlying factor of reading acquisition (
Even though the relation of phonemic awareness and word reading skills is well documented in the literature (
Regarding number transcoding and phonological processing skills, there is even less investigation. Even though mathematical and reading/spelling disabilities have a high comorbidity rate (
The association between spelling and word reading to mathematics disabilities depends on the cutoff criteria used to define them. Most studies have investigate associations between word reading and arithmetic, especially in dyslexic samples (
The use of intelligence as a covariate in studies of children with learning disabilities has been criticized (
Another line of argumentation is also defensible. Estimates of general cognitive ability have been repeatedly found to be among the best predictors of school achievement and other psychosocial outcomes (
Mastery of word and number processing at the lexical level is an enormous task for children at early school age. There is evidence, for example, that mastery over Arabic number dictation is reached only after 3–4 years of schooling in typically developing children (
Early school abilities related to word and number reading and writing may depend on both domain-general and domain-specific cognitive abilities. Sources of influence are both shared and non-shared across codes and tasks. Phonological and visuospatial WM tasks could be uniquely associated, respectively, to verbal lexical and numerical tasks. But these effects disappear or are attenuated when general non-verbal intelligence is covaried (the zero-order correlation between number writing and digit span is
These findings suggest that phonemic awareness may be considered as a domain-specific cognitive mechanism which is strongly associated to reading and writing of both numbers and words. From these results, one can infer that phonemic awareness is a mechanism shared by numerical and verbal domains, which might also be a candidate associated to the high comorbidity rate between dyslexia and dyscalculia. It is important to note that this specific mechanism is related to number and word reading and writing and should, therefore, be taken into account in intervention models.
All authors have contributed significantly and all of us agree with the content of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
This work was supported by grants from CAPES/DAAD Probral Program, Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, 307006/2008-5, 401232/2009-3 and Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG, APQ-02755-SHA, APQ-03289-10, FAPEMIG, APQ-02953-14, APQ-03642-12). VGH is supported by a CNPq fellowship (308157/2011-7, 308267/2014-1). GW is supported by an FWF research project (No. P22577).