Automaticity and written language: An investigation of the written language performance of second- and fourth-grade students with and without learning disabilities

Date of Completion

January 1994


Education, Elementary|Education, Special




The written language performance of learning disabled students has been found to be quantitatively and qualitatively inferior of that of their nondisabled peers. Most research has been directed at describing the deficiencies in the written language compositions of students with learning disabilities, rather than investigating possible underlying cognitive processes as has been done in the reading and mathematics domains. Although some researchers have proposed that lower level skills such as handwriting, spelling, and sentence generation must be developed to an automatic level in order for higher level processes, such as theme development and text cohesion, to occur, these theories have not been empirically tested. This study was designed to investigate the relationship between fast, accurate execution of lower level skills ("automaticity") with the production of an integrated writing task (story composition).^ Two hundred seventeen learning disabled and nondisabled students in second and fourth grade classrooms in Connecticut served as subjects for the study. Subjects completed five tasks that involved automaticity: letter identification, word recognition, writing the alphabet, word fluency, and sentence fluency. Subjects also produced a narrative story composition that was rated for fluency, vocabulary, theme elements, spelling, and cohesion. An overall story quality score was also computed, and two standardized tests of reading and spelling skills were administered to provide measures of achievement.^ Results of the MANCOVA and planned Stepdown Analyses indicated differences between grade level groups and disability groups, even after initial adjustment for differences in reading and spelling achievement. Fourth grade subjects were superior to second grade subjects; and nondisabled subjects were superior to learning disabled subjects on several of the measures of automaticity and story composition variables. Automatic word recognition and spelling were particularly sensitive to differences between groups.^ Discriminant Function Analysis indicated that group membership was reliably predicted using four of the five automaticity variables and five of the story composition variables. Correlation matrices indicated strong relationships between automaticity and overall quality of the story for all groups. Findings from this study suggest that the poor quality of LD students' written language may be related to underlying difficulties with lower level skills. ^