What counts as an error in nonword reading? A methods study of acquired phonological dyslexia

Abstract

Acquired phonological dyslexia (PD) is typically characterised by impaired nonword reading, but defining what constitutes a nonword reading “error” is challenging due to variability in how even healthy readers pronounce nonwords. This study examined different methods for classifying such errors using data from younger adults (𝑛 = 41), older adults (𝑛 = 25), and a case series of individuals with suspected acquired PD (𝑛 = 6). Three scoring approaches were compared: (1) a mode-response accuracy method, which matches responses to the most common pronunciation; (2) an average response-overlap method, which assesses similarity to the full distribution of control responses; and (3) a singularly distinct response count, which identifies responses not produced by any other participant. Results showed that older adults exhibited greater variability than younger adults across all measures, suggesting that age significantly affects nonword reading performance. This highlights an important methodological issue: using younger normative data to assess older individuals may overestimate impairment. In the patient group, all three scoring methods broadly agreed in ranking severity across cases. However, the singularly distinct response metric provided clearer separation between patients and controls and captured a wider range of impairment severity. It also revealed that unique responses are relatively common even among typical readers, an underexplored finding. Overall, the study supports the importance of age-appropriate norms and more nuanced, response-level scoring methods in assessing acquired PD, with singularly distinct responses offering a potentially sensitive marker of disruption.