Saturday, November 12, 2005

Ga, Glr, Gs and early reading - relative contributions and shared prediction variance

Although the conclusions of this study must be tempered by the small sample size, the study provides some interesting conclusions regarding the potential shared variance between some of the “hot” abilities that have been identified in the early reading literature. Of particular interest is the conclusion that the effect of RAN (Glr-NA) may be mediated via phonological processing (Ga-PC). The relative contribution of such CHC conceptualized abilities as Glr-NA, Gs, and Ga-PC to the prediction of reading are interesting to ponder.

Bowey, J., McGuigan, M. & Ruschena, A. (2005) On the association between serial naming speed for letters and digits and word-reading skill: Towards a developmental account Journal of Research in Reading, 28 (4) 400–422. [Click here to view entire article]

Abstract [Note - CHC ability codes inserted by blogmaster]
  • The current study examined several alternative explanations of the association betwee serial naming speed within fourth-grade children by determining the extent to which the association between word reading [Grw-RD]and naming speed [Glr-NA]for letters and numbers is mediated by global processing speed (Gs], alphanumeric symbol processing efficiency and phonological processing ability [Ga-PC]. Children were given multiple measures of key constructs, i.e. word-level reading, serial naming of both alphanumeric and non-alphanumeric items, phonological processing ability, articulation rate and global processing speed. The robust association between alphanumeric naming speed and reading within fourth-grade children was largely mediated by phonological processing ability. Markedly different patterns of results were observed for naming speed for letters and digits and naming speed for colours and pictures in children of this age. Relative to the latter, alphanumeric naming speed better assesses an underlying phonological processing ability that is common to word-reading ability. We argue that item identification processes contribute little to individual differences in alphanumeric naming speed within relatively proficient readers and that the extent to which alphanumeric naming speed primarily reflects phonological processing is likely to vary with the level of overlearning of letters and numbers and their names.

Relative importance of Ga
  • Overall, our findings suggest that phonological processing ability largely mediates the association between alphanumeric naming speed and reading within children who are beyond the earliest stages of reading acquisition. In our fourth graders, phonological processing ability explained 61% of variance in word reading. The major source of the variance shared by naming speed and word reading was phonological processing ability. Of the 21% of variance shared by alphanumeric naming speed and word reading, only 2% was independent of phonological processing ability; of the 8% of variance shared by non-symbol naming speed and word reading, only 2% was independent of phonological processing ability. However, alphanumeric naming speed shared 19% of variance with phonological processing ability but non-symbol naming speed shared only 6%. Given these findings, it follows that one or more components of the non-symbol naming task that contribute substantially to variance in naming speed are not shared with word reading, phonological processing ability or naming speed for letters and digits.
Relative importance of Gs
  • Replicating previous work, we found that global processing speed was associated with word-reading ability. In our fourth-grade children, global processing speed explained 13% of variance inword reading However, although global processing speed largely mediated the association between non-symbol naming speed and word-reading skill, it did notmediate the association between word-reading skill and serial naming speed for lettersand digits; with global processing speed effects controlled, alphanumeric naming speed still explained 12% of additional variance in reading.
Relative importance of different Glr-NA measures/constructs
  • Our findings revealed the clear need to distinguish between naming speed for letters and digits and naming speed for colours and pictures. Composite measures of alphanumeric and non-symbol naming speed were each significantly correlated with word-level reading, but the association was considerably stronger for alphanumeric naming speed. While naming speed for letters and digits shared 21% of variance with word reading, naming speed for colours and pictures of common objects shared only 8%. The tasks of rapidly naming letters and digits and colours and pictures obviously include the same sub-processes, and our composite measures of these tasks shared 41% of variance. Relative to rapidly naming non-symbol items, alphanumeric naming speed better assesses one or more underlying abilities or additional processing components common to word-reading skill. One candidate is some aspect of automatised orthographic processing (Bowers & Wolf, 1993). According to this view, alphanumeric naming speed predicts variation in word-reading skill independently of phonological processing ability. Indeed, this claim comprises a central argument in the rationale underlying the orthographic processing efficiency account (Bowers & Wolf, 1993; Bowers et al., 1994). Alphanumeric naming speed tasks require children to identify and name a continuous series of letters or single-digit numbers as fast as they can. Non-symbol naming speed tasks differ only in that children are required to name colour patches or pictures of familiar objects. Serial alphanumeric and non-symbol naming tasks thus differ primarily in the familiarity of the items to be identified and named. Letters and numbers are encountered far more frequently than line drawings of objects and within standard fonts are probably identified on the basis of a constrained set of features, making for automatised identification within children of this level of reading development. Our findings revealed quite different patterns of results for naming speed for letters and digits and naming speed for colours and pictures in children of this age. Relative to the latter, alphanumeric naming speed better assesses an underlying phonological processing ability that is common to word-reading ability. The variance shared by alphanumeric naming speed and word reading did not appear to reflect an independent orthographic processing ability. We argue that item identification processes contribute little to individual differences in alphanumeric naming speed within relatively proficient readers. The extent to which alphanumeric naming speed primarily reflects phonological processing is likely to vary with the level of over-learning of letters and numbers and their names, so that predictably different outcomes may be observed in beginning readers and within severely disabled readers.

No comments: