Predicting individual differences in reading, spelling and maths in a sample of typically developing children: A study in the perspective of comorbidity

Abstract

We examined reading, spelling, and mathematical skills in an unselected group of 129 Italian fifth graders by testing various cognitive predictors for each behaviour. As dependent variables, we measured performance in behaviours with a clear functional value in everyday life, such as reading a text, spelling under dictation and doing mental and written computations. As predictors, we selected cognitive dimensions having an explicit relation with the target behaviour (called proximal predictors), and prepared various tests in order to select which task had the best predictive power on each behaviour. The aim was to develop a model of proximal predictors of reading (speed and accuracy), spelling (accuracy) and maths (speed and accuracy) characterized by efficacy also in comparison to the prediction based on general cognitive factors (i.e., short-term memory, phonemic verbal fluency, visual perceptual speed, and non-verbal intelligence) and parsimony, pinpointing the role of both common and unique predictors as envisaged in the general perspective of co-morbidity. With one exception (reading accuracy), the proximal predictors models (based on communality analyses) explained a sizeable amount of variance, ranging from 27.5% in the case of calculation (accuracy) to 48.7% of reading (fluency). Models based on general cognitive factors also accounted for some variance (ranging from 6.5% in the case of spelling to 19.5% in the case of reading fluency) but this was appreciably less than that explained by models based on the hypothesized proximal predictors. In general, results confirmed the efficacy of proximal models in predicting reading, spelling and maths although they offered only limited support for common predictors across different learning skills; namely, performance in the Orthographic Decision test entered as a predictor of both reading and spelling indicating that a single orthographic lexicon may account for performance in reading and spelling. Possible lines of research to expand on this approach are illustrated.

Testing the Specificity of Predictors of Reading, Spelling and Maths: A New Model of the Association Among Learning Skills Based on Competence, Performance and Acquisition

“…In the present report, as well as a previous companion one ( Zoccolotti et al, 2020 ), we aimed to develop a unitary model to uncover the unique and shared influences of predictors for reading, spelling and maths skills. To this aim, we examined these performances in an unselected group of fifth grade Italian children.…”

“…The present report is strictly linked to a previous one based on the same dataset ( Zoccolotti et al, 2020 ), which presented the following main features. First, as target dependent behaviors we selected ecological tasks, such as reading a text passage, spelling a passage under dictation, and making calculations.…”

“…Second, we used a “proximal” approach, that is we formulated explicit causal relationships between predictors (i.e., cognitive antecedent of the specific behavior under scrutiny) and target dependent measures (i.e., reading, spelling or maths). Third, based on the relevant literature, predictors of reading, spelling and maths were selected pointing to both efficacy and parsimony (details of such a selection are given in Zoccolotti et al, 2020 ). Fourth, as a control, we tested the possible predictive power of general cognitive dimensions (i.e., measures of short-term memory, phonemic verbal fluency, visual perceptual speed, and non-verbal intelligence).…”

“…Part A) Predictors in the original models of reading, spelling, and maths and in the general cognitive factors model (fromZoccolotti et al, 2020). Part B) Percentage of total variance explained by different models.…”

“… In a previous study ( Zoccolotti et al, 2020 ) we examined reading, spelling, and maths skills in an unselected group of 129 Italian children attending fifth grade by testing various cognitive predictors; results showed a high degree of predictors’ selectivity for each of these three behaviors. In the present study, we focused on the specificity of the predictors by performing cross-analyses on the same dataset; i.e., we predicted spelling and maths skills based on reading predictors, reading based on maths predictors and so on.…”