What Reading-Level Match Design Reveals about Specific Reading Disability in a Transparent Orthography and How Much We Can Trust It

Purpose To assess strengths and weaknesses of the reading level (RL) match approach and its potential to generate insights regarding the cognitive foundations of reading ability and disability. Method We applied RL-match design to a sample of 2(nd) - 6(th) graders reading a consistent orthography, Russian, using an "extreme phenotype" approach. Readers with suspected specific reading disability (sSRD, n = 538) and high-performing readers (HPR; n = 806) were matched via propensity Scores, using IQ and each of the alternatives: accuracy of word decoding, pseudoword decoding, word unitization, or paragraph reading fluency. In each case, two groups were compared on the remaining literacy tasks as well as phonological processing, orthographic processing, and rapid serial naming. Results When matched on word or pseudoword decoding (288 and 313 pairs, respectively), readers with sSRD and HPR differed on all remaining indicators. When matched on word unitization (173 pairs), the differences disappeared or had substantially diminished effect sizes. When matched on paragraph reading fluency (57 pairs), no significant differences remained. Thus, none of the componential skills appeared antecedent to the observed difficulties assessed via the number of correctly orally read words per minute. However, certain inherent limitations of RL-match design preclude us from considering this to be a definitive outcome.