Which tasks best discriminate between dyslexic university students and controls in a transparent language?

The need for a battery for testing adult dyslexia, and especially university students, is being increasingly recognized in view of the increased number of adult requests for a dyslexia examination in relation to both assistance and protection from discrimination. The present study examines the discriminative validity of a battery we have developed-the Battery for the Assessment of Reading and Writing in Adulthood-through comparison of the performance of 24 university students with a history of severe developmental dyslexia and 99 controls. All the reading, writing, lexical decision and spelling tasks of the battery, except omissions in a lexical decision task and reading comprehension, showed a good discriminatory power. In addition, use of just two of these tasks (fluency in reading a text and spelling under articulatory suppression) gave 87% sensitivity and 97% specificity. Our results confirm that in transparent languages, measures of phonological automaticity are the best indexes of reading decoding competence, particularly in adults.