Analyzing Spelling Errors by Linguistic Features among Children with Learning Disabilities

In order to spell fluently and accurately, phonology, orthography, and morphology must be integrated and stored into long term memory (Berninger & Richards, in press; Berninger, Nagy, Tanimoto, Thompson, Abbott, 2015). Children with dysgraphia, dyslexia, and OWL-LD have specific deficits in linguistic processing that impede the cross-mapping of these linguistic elements. This study analyzes the frequency and nature of spelling errors produced by children with dysgraphia, dyslexia, and OWL-LD during an academic writing task in order to determine if known deficits in linguistic processing affect the type and severity of spelling errors made by these children. The present study analyzed error severity and frequency of spelling errors produced by children with dysgraphia (n=13), dyslexia (n=17), or OWL-LD (n=5) during the academic writing tasks obtained in the Berninger et al. (2015) study. In the previous study, students read or listened to computerized lessons about basic mathematical concepts and then typed summaries of what they learned. For the current study, all spelling errors made during the typed summary writing tasks were extracted and analyzed using the Phonological, Orthographic, Morphological Assessment of Spelling (POMAS) and then recoded with POMplexity (a measure of error severity) to determine the severity and frequency of spelling errors made in the linguistic categories of phonology, orthography, and morphology. Results indicated that the students did not differ in error severity by diagnostic category. However, a qualitative analysis using the POMAS revealed that children from different diagnostic categories produced different types of errors. With respect to error frequency, only students with dysgraphia made significantly fewer errors than students with OWL-LD, and all participants, regardless of diagnostic category produced more errors in typed summaries following the reading condition. These results are consistent with previous research indicating that children with learning disabilities do not produce deviant spelling errors when compared to typically-developing, age-matched peers or typically-developing, spelling-matched peers (Silliman, Bahr, and Peters, 2006, among others). The current results demonstrate that the spelling errors of children with learning disabilities reflect the expected linguistic breakdowns in cross-code mapping, and that children with learning disabilities may display these spelling deficits beyond an appropriate age