Relation of phonological and orthographic processing to early reading: Comparing two approaches to regression-based, reading-level-match designs.

K. E. Stanovich and L. Siegel (1994) introduced regression-based logic to the reading-level-match design by statistically matching children with reading disabilities, with and without discrepancies in IQ, to normal-reading children on the basis of grade-adjusted decoding scores. The authors replicated this approach but contrasted it with statistical matches using w scores, which are Rasch-scaled decoding scores based on a common metric regardless of age or grade. No differences were found in cognitive skills between children whose reading performance was discrepant and not discrepant with IQ, regardless of whether age-adjusted decoding scores or w scores were used. Matching on w scores did not result in the phonological and orthographic tradeoffs seen when standardized scores were used. The orthographic-decoding relationship was nonlinear, with little functional relation between the skills at low levels of decoding. These results question the conclusion that orthographic skills are compensatory for reading-disabled children.