The Contribution of Set for Variability in Reading: Does it Varies across the Word Reading Distribution?

The Contribution of Set for Variability in Reading: Does it Varies across the Word Reading Distribution?

First Author: Nuria Gutiérrez -- Florida State University
Additional authors/chairs: 
Alexandra Himelhoch; Valeria Rigobon; Ashley Edwards; Laura Steacy ; Donald Compton
Keywords: Word reading, Reading disability, Quantile regression
Abstract / Summary: 

Purpose. The ability to determine the correct pronunciation of approximations to spoken words (i.e., set for variability [SfV]) has shown to have an essential role in word recognition, both for regular and irregular words. This study aims to answer two research questions: (1) What is the contribution of SfV in word reading after controlling for other reading-related variables? (2) Does this contribution vary across the distribution of word recognition?
Method. A sample of approximately 425 students from grades 2–5 were assessed from private and public schools. Participants’ mean age was 9.42 years. Multivariate quantile regression was used to evaluate the extent to which SfV had a differential explanatory effect across different quantiles of the word reading distribution. Other reading-related skills (e.g., phonological awareness, vocabulary, RAN, attention) were also introduced in the regression model.
Results. Preliminary analysis suggested that SfV explains a unique proportion of the variance in word reading. Furthermore, quantile regression analyses also revealed that the contribution of SfV was larger in the higher quantiles of the outcome distribution.
Discussion. Word recognition in an opaque orthography can be challenging, especially for those students presenting reading difficulties. Results suggest that SfV is an essential piece in the process that “cleans up” the discrepancies between the regularized (i.e., using decoding strategies) and the correct pronunciation of the word. Implications for intervention and assessment will be discussed.