Efficient word reading: Automaticity of print-related skills indexed by rapid automatized naming through cusp-catastrophe modeling

ABSTRACT

The study explored the moderating role of rapid automatized naming (RAN) in reading achievement through a cusp-catastrophe model grounded on nonlinear dynamic systems theory. Data were obtained from a community sample of 496 second through fourth graders who were followed longitudinally over 2 years and split into 2 random subsamples (validation and cross-validation groups). Results verified the superiority of the cusp-catastrophe over linear and logistic models and established RAN-digits performance as a significant bifurcation factor in both concurrent and longitudinal prediction models. These findings suggested that reduced serial naming speed below a critical level was associated with significantly reduced predictability of word reading efficiency from pseudoword decoding ability. Results were tentatively interpreted as documenting the importance of highly automatized coordination of reading-related component processes, indexed by RAN performance, for word reading efficiency during the early and middle stages of reading acquisition.

Notes

¹ Grasman et al. (2009) defined the dependent variable as a “smoothed approximation” of an actual variable and similarly the parameters a and b as canonical approximations of the control variables.

² For which Gignac and Watkins (2013) recommended that difference values greater than 10 points are indicative of “very strong” evidence in favor of the model with the lower value.

³ In regression analysis the −2 times the log likelihood is replaced by N times the log of the variance of the residuals.