This study examined longitudinally the double-deficit hypothesis in Greek, an orthographically consistent language, following a group of children from kindergarten to Grade 2. Four groups were formed on the basis of two composite scores of phonological and naming-speed criterion measures: a double-deficit group (DD; n = 17), a phonological deficit group (PD; n = 33), a naming deficit group (ND; n = 33), and a control group exhibiting no deficits (CnD; n = 159). The four groups were identified in Grade 1, and they were compared retrospectively in kindergarten only on the criterion measures, and in Grades 1 and 2 on measures of word-reading fluency and accuracy, orthographic processing, and passage comprehension. The effects of verbal and nonverbal ability, age, gender, and parental education were controlled among the groups. Results showed that the DD group exhibited greater dysfunction in reading and orthographic processing compared to the single-deficit and CnD groups. Also, although the three deficit groups were not easily differentiated in kindergarten, their differences were maximized in Grade 1 and retained in Grade 2. The type and severity of reading deficits found in the ND group were mostly associated with naming speed at both the word- and text-reading levels, deficits that persisted across development. The PD group showed mostly deficient orthographic and poor decoding skills that improved across development. Implications of the findings for the double-deficit hypothesis in languages with transparent orthographies are discussed.
We examined the prominent theoretical explanations of the RAN-reading relationship in a relatively transparent language (Greek) in a sample of children (n= 286) followed from Grade 1 to Grade 2. Specifically, we tested the fit of eight different models, as defined by the type of reading performance predicted (oral vs. silent word reading fluency), the type of RAN tasks (non-alphanumeric vs. alphanumeric), and the RAN effects (direct vs. indirect). Working memory, attention, processing speed, and motor skills were used as “common cause” variables predicting both RAN and reading fluency and phonological awareness and orthographic processing were used as mediators of RAN’s effects on reading fluency. The findings of both concurrent and longitudinal analyses indicated that RAN is a unique predictor of oral reading fluency, but not silent reading fluency. Using alphanumeric or non-alphanumeric RAN did not particularly affect the RAN-reading relationship. Both phonological awareness and orthographic processing partly mediated RAN’s effects on reading fluency. Theoretical implications of these findings are discussed.