Characterization of reading errors in languages with different orthographic regularity: an Italian–English comparison

The study examined whether a classification of errors based on Hendriks and Kolk's (1997) proposal would effectively characterize the reading profile of children learning two orthographies varying for regularity, such as Italian and English. The study considered both an age-match and a grade-match comparison. Offline analysis of error production was carried out for two lists of stimuli: List 1 including regular words varying for frequency and matched non-words and List 2 including low-frequency words varying for regularity. In List 1, Italian-reading children made more multiple attempts characterized by a slow and progressive approach to the target (sounding-out behavior) than English-reading children, while the latter made relatively more word substitutions and non-word lexicalizations. As for List 2, Italian-reading children made relatively more multiple attempts and progressive approaches to the target compared to the English-reading children (with more sounding-out behaviors and syllabications), while the opposite occurred for phonological-visual errors, word substitutions, morphological, and semantic errors. Both groups showed a high proportion of phonological-visual and regularization errors (stress assignment in the case of Italian-reading children). Overall, the use of an error coding system specifically tuned to the characteristics of the orthographies investigated allowed a more comprehensive identification of reading difficulties which allowed the different strategies used by children of different languages to emerge more clearly (more reliance on sub-lexical routines in Italian readers and on lexical routines in English readers). These results call for more attention to error patterns in the identification of reading difficulties in children of different languages including those learning a transparent orthography where error analyses have largely been ignored.