MacArthur, C. A., & Graham, S. (1987). Learning disabled students’ composing under three methods of text production: Handwriting, word processing, and dictation . The Journal of Special Education , 21 (3), 22–42. https://doi.org/10.1177/002246698702100304

Each student composed three stories, one using each method of text production: handwriting, word processing, dictation. Each story was composed in response to a colored picture.

Participants included 11 fifth- and sixth-grade students (six male and five female, six black and five white). Each student had been identified as having a learning disability and attended a special education resource room program for approximately one hour a day. Their mean age was 143.6 months. Their mean performance on the Test of Written Language was 82.59. Nine participants wrote with word processors in regular class day, five used word processors at home, and all but two used a word processor at least once per week for 30 minutes.

Stories that students wrote were rated according to five factors. Language Complexity was measured by number of words, average T-unit length, corrected type-token ratio, number of different words divided by the square root of 2 times the total words, proportion of mature words, and proportion of grammatical errors to total words. Mechanical Errors included spelling, capitalization, and punctuation. Quality and Story Structure involved a holistic evaluation procedure on an 8 point scale, with schematic structure of eight story grammar elements -- main character, locale, time, starter event, goal, action, ending, and reaction. Time and Rate Measures were pre-writing time, composing time, and composing rate. Revisions included syntactic level of the change, such as word or sentence, and the type of operation, such as addition or deletion; in all, five levels were analyzed -- surface, word, multi-word, T-unit, and multi-T-unit.

The results demonstrate that dictation differed considerably from both handwriting and word processing. Dictated stories were significantly longer and of higher quality. They also had fewer grammatical errors. For students with LD, the mechanical and conventional demands of producing text appeared to interfere with the fluency and quality of written expression. When these demands were removed via dictation, students with LD composed more fluently and with better results. Dictation was approximately 9 times faster than handwriting and 20 times faster than word processing. In contrast to the observed differences between dictated and written stories, no significant differences between handwriting and word processing were found on any of the product measures. Handwritten and word processed stories did not differ on length, quality, story structure, mechanical or grammatical errors, vocabulary, or average T-unit length. Although word processing was less than half as fast as handwriting, the overall amount of revision was similar for handwriting and word processing, as was the syntactic level of the revisions.