Orthographic depth and developmental dyslexia: a meta-analytic study

This study conducted a meta-analysis of 79 peer-reviewed articles published between 2013 and 2018 to investigate the cognitive and behavioral manifestations of developmental dyslexia (DD). The researchers sought to determine which reading and cognitive deficits are universal across different European languages and which are moderated by orthographic depth (the consistency of grapheme-to-phoneme mapping) and the developmental stage (children vs. adults). By categorizing languages into shallow and deep orthographies, the authors assessed performance differences between dyslexic and control groups in word reading, non-lexical decoding, phonological awareness (PA), rapid automatized naming (RAN), working memory, and nonverbal reasoning.

The study identified a clear distinction between reading accuracy and reading fluency. While word reading accuracy showed a significant interaction between age and orthographic depth—suggesting that the gap between dyslexic and typical readers narrows differently depending on the language's transparency—reading fluency emerged as a universal, stable index of DD. Regardless of the language or the reader's age, dyslexic individuals consistently demonstrated impaired reading speed. Furthermore, cognitive markers including phonological awareness, rapid automatized naming, and short-term working memory were found to be impaired in dyslexic readers across all ages and orthographies, supporting the phonological theory and the double-deficit hypothesis of dyslexia.

The authors argue that clinical practice should prioritize reading fluency over accuracy-based assessments, as accuracy can be artificially inflated by compensation strategies or orthographic transparency. They recommend that clinicians routinely include PA and RAN assessments in their diagnostic batteries. Additionally, the study suggests that nonverbal reasoning remains a valid criterion for distinguishing DD from general intellectual impairments, as the observed differences in reasoning between dyslexic and control groups were minimal and did not predict reading outcomes.