Moving
beyond the monosyllable in models of skilled reading:
Mega-study
of disyllabic nonword reading
The present paper reports a
large-scale study in which 41 participants read aloud 915 disyllabic nonwords,
yielding a total of around 37,000 responses. We investigated the cues to stress
assignment and the factors that influence pronunciation variability and reading
latencies in the English language. We also compared human reading performance to
the reading performance of computational models of reading that adopt
rule-based and statisticallearning approaches to explaining disyllabic reading.
Our findings provided support for the latter approach, although we identified
important deficiencies with the most fully developed model of this type. The
findings from the present study make a critical theoretical contribution to our
understanding of skilled adult reading. Further, this dataset provides the
first normative nonword corpus for British English and is the largest database
of its kind for any language, thus being critical for evaluating generalization
in models of reading as they advance into the disyllabic domain. Finally, our
findings have significant applied implications for the development of
evidence-based strategies for literacy education and the clinical diagnosis and
treatment of reading impairments.
No comments:
Post a Comment