Over the last several years, researchers have focused their attention on deficits in the processes underlying naming speed as another possible source of reading difficulties (Bowers and Wolf, 1993; Wolf and Bowers, 1999; 2000). The rationale behind the research aimed at investigating other possible factors beyond phonology that influence reading difficulties, was based on the fact that many reading disabled children who eventually mastered grapheme-phoneme correspondence rules continued to exhibit deficits in reading speed and reading comprehension. Gradually an alternative conceptualisation of reading difficulties has been developed suggesting that naming speed deficit provides a source of reading difficulties that is largely independent of phonological awareness difficulties. In the case where the two deficits (the phonological awareness deficit and naming speed deficit) co-exist more severe reading difficulties are observed. This condition has been termed as a double deficit.

The importance of studying naming speed deficit in transparent languages is even greater since the phonological demands in these languages are decreased and, therefore, it is easier to examine the influence of naming speed on reading development and/or reading difficulties. One would hypothesize that in a transparent language where the demands for phonological analysis are reduced naming speed deficits would be the dominant diagnostic index for reading difficulties.

The aim of this study was to investigate whether there is a naming speed deficit as an independent core feature of reading difficulties or an associate feature of a phonological deficit in Grade 1 Greek speaking children.

Method

A population of 276 grade 1 children, coming from 42 urban schools that traditionally collaborate with the University of Cyprus for research and training purposes, participated in this study. The following tests were administered to form the participating deficit groups: Rhyming, Rhyme Generation, Syllable-splitting, Onset-Rime, Initial Syllable Oddity Task, Final Syllable Oddity task (tapping broad phonological awareness), Phoneme Elision, Sound Isolation, Blending, and Initial Sound Alliteration (tapping narrow phonological awareness) and Naming of Pictures, Colors, Letters and Digits. Those tests that revealed significant high correlations (r > .40) with two composite scores for reading speed and fluency of real and pseudoword reading were used to form the participating deficit groups. By including only the dominating phonological and naming tasks in the subsequent analysis, it was ensured that deficit group assignment would be relatively robust. Reading speed and word accuracy of real and nonwords, spelling, and passage comprehension were used as the dependent variables. Verbal and non-verbal ability as well as parental education level and age were used as group matching variables. To test the Double Deficit Hypothesis (DDH) four groups were formed: a Double Deficit group (DD; n=22), a Phonological deficit group (PD; n=32), a Naming deficit group (ND; n=33), and a Control group exhibiting no deficits (CnD; n=189). The bottom 20^th quartile of the sample mean on the composite scores of five phonological measures (Phoneme Elision, Sound Isolation, Blending, Initial Sound Alliteration, and Rhyming) and of two rapid-naming measures (naming of letters and digits) was used as criterion to form the three deficit groups.

Results

Results revealed a significant overall effect indicating that the four groups performed differently on the dependent measures [multivariate Wilks L = .568, F(8,265) = 6.91, p < 001]. Indeed, the three deficit groups performed significantly lower than the CnD group in all the dependent measures. The same results emerged when the DD group was compared to the single deficit groups, based on post-hoc planned comparisons. Surprisingly enough, the only group difference that was observed between PD and ND groups was on a spelling task, the orthographic choice, a finding that was not expected on the basis of previous research testing the DDH in languages with salient orthographies (see for e.g., Wimmer et al., 2000).

Discussion

The results of the present study challenge the double deficit hypothesis in Greek in early years. All three deficit groups (PD, ND, DD) showed the same pattern of reading performance compared to the CnD group. Therefore, it is possible to suggest that naming speed does not independently influences reading difficulties in a salient orthography, as expected, but, instead, it might share some variance associated with phonological awareness.

Overall, the performance of the three deficit groups on reading accuracy and speed suggests that rapid naming is largely associated with phonological processes implying the existence of a phonological deficit in symbol processing speed. That is, phonological representations such as mental representations of word pronunciation or syllable and/or sound manipulations are less accessible in all three deficit groups and this could be the reason why it takes them longer to activate these representations in RAN tasks or, conversely, to perform successfully demanding phonological tasks.