Two Decades of Reading Performance at the Alouette Test: Descriptive Data and Dyslexia Screening Properties in French Primary School Children

Background: The Alouette test is widely used in France to assess reading performance and screen for dyslexia. Its low semantic predictability text limits contextual compensation and emphasizes word identification. However, its 2005 standardization relied on small grade-level samples, evidence on its screening properties in school-age children remains limited, and temporal changes in performance are poorly documented. Methods: We pooled 5,118 Alouette assessment records from typically developing French-speaking children in Grades 1–5, collected across research projects between 2004 and 2025. We examined grade and sex differences and associations with school socioeconomic status. Screening properties were evaluated in 744 children with a prior diagnosis of developmental dyslexia and 744 matched controls. Temporal associations were examined using regression models accounting for grade, sex, and school socioeconomic status, with school-level clustering addressed where appropriate. Sensitivity analyses examined age and school composition, and exploratory models compared gradual trends with change around 2018, when French educational reforms placed renewed emphasis on foundational reading skills and structured phonics. Results: Performance improved across grades on all indices except errors, with no substantial sex differences. Associations with school socioeconomic status were stronger in the upper grades. Children with dyslexia performed more poorly on all indices. The areas under the Receiver Operating Characteristic (ROC) curves were .861 for reading accuracy and .902 for reading efficiency. At grade-specific 10th-percentile thresholds, reading efficiency yielded 76.3% sensitivity and 88.6% specificity. Across assessment years, reading efficiency increased by about 0.6% per year on average, whereas accuracy declined slightly, mainly in Grades 1 and 2. Exploratory models favored a difference of approximately 9% in reading efficiency between observations collected before and after 2018 over a single linear trend, although patterns varied across indices.Conclusions: This study provides extensive descriptive data on Alouette performance and evidence of its potential for dyslexia screening. The proposed thresholds require independent validation before routine clinical use. Retrospective sampling and discontinuous temporal coverage limit generalizability, and the observed post-2018 difference cannot be attributed with certainty to education reforms.