Presented at the 20th Annual Conference of the Society for the Scientific Study of Reading; Hong-Kong, July 10–13 2013.

Purpose: Serial rapid automatized naming (RAN) is significantly related to reading fluency across ages and orthographies. Naming of isolated words, thought to assess lexical access, is relatively less strongly related to fluency. We aimed to explore this difference in order to examine the structure of the RAN-reading relationship. We were particularly interested in developmental differences in the underlying pattern of interrelations.
Method: 107 Greek children from Grade 2 and 107 from Grade 6 were tested with discrete and serial naming of digits, objects, and words in 50-item arrays.
Results: The correlation between discrete and serial word reading was very high in Grade 2 but only moderate in Grade 6. In confirmatory factor analysis, a reading-naming latent structure fit the Grade 2 data well; in contrast, a serial-discrete structure fit the Grade 6 data. Discrete RAN failed to account for any word fluency variance in either grade, whereas serial RAN was a significant unique predictor in both grades.
Conclusions: The superficial longitudinal stability of RAN-reading correlations belies vastly different patterns of relations, indicative of changes in developing cognitive processes. Word fluency tasks in Grade 2 are apparently accomplished as a series of isolated individual word naming trials. In contrast, specifically serial procedures are applied in Grade 6. This must be achieved via simultaneous processing of multiple individual words at successive levels, a feat that requires endogenous (not stimulus-driven) control and monitoring of cognitive cascades. We hypothesize that RAN indexes the development of this executive control, concurrently, and the requisite propensity, longitudinally.