ABSTRACT
Abstract Phonological awareness is one of the most important predictors of reading. However, there is still controversy concerning its dimensionality. This study evaluated the dimensionality of phonological awareness among Brazilian Portuguese-speaking children. A total of 212 children performed six phonological awareness tasks in the last year of kindergarten. Of those children, 177 performed the same tasks when they were in the first grade. The phonological awareness measures differed in both their cognitive demand (detection, blending, segmentation, and elision) and the phonological unit involved (rhyme, syllable, and phoneme). Confirmatory factor analyzes were employed to test several models of phonological awareness dimensionality. The results indicated that the best model was an oblique model of phonological units with two correlated latent factors: phonemic awareness and supraphonemic awareness. This model presented the best fit to the data both in kindergarten and in the first grade. In addition, supraphonemic awareness in the kindergarten predicted phoneme awareness in the first grade; however, phonemic awareness in the kindergarten did not predict supraphonemic awareness in the first grade. These results are compatible with phonological awareness developing from larger phonological units (e.g., syllables) to small phonological units (e.g., phonemes) and the reciprocal relationship between phonological awareness and reading. From a theoretical point of view, these results also suggest that phonological awareness is a one-dimensional construct that can be evaluated by tests employing different phonological units (e.g., syllables, rhymes, phonemes).
Subject(s)
AwarenessABSTRACT
O presente estudo investigou a natureza das tarefas de nomeação seriada rápida, considerando três hipóteses sobre o que elas representariam: (1) medidas da recuperação de códigos fonológicos da memória de longo prazo; (2) medidas da velocidade de processamento geral; e (3) medidas de processos relacionados ao reconhecimento de padrões visuais. Participaram da pesquisa 174 crianças brasileiras cursando o 5º ano do Ensino Fundamental em escolas particulares. Análises fatoriais confirmatórias indicaram que nenhuma das três hipóteses se mostrou consistente com os dados. Porém, em um modelo de regressão estrutural construído para avaliar a contribuição desses fatores para a nomeação seriada rápida, o único que contribuiu significativamente foi o fator 'velocidade de processamento geral'...
The present study aimed to investigate the nature of rapid automatized naming tasks, considering three hypotheses about what they represent: (1) measures of retrieval of phonological codes from a long-term store; (2) measures of general processing speed; and (3) measures of visual pattern recognition processes. The participants were 174 children from the 5th grade of Brazilian private schools. Confirmatory factor analyses indicated that none of the hypotheses were consistent with the data. However, in a structural regression model the factor 'general processing speed' was the only one that contributed significantly to rapid automatized naming...