MacArthur, C. A., & Cavalier, A. R. (2004). Dictation and speech recognition technology as test accommodations . Exceptional Children , 71 (1), 43–58. https://doi.org/10.1177/001440290407100103

This study addressed the feasibility and validity of dictation using speech recognition software (Dragon Naturally Speaking, Version 4) and dictation to a scribe as test accommodations for students with learning disabilities.

The participants included 31 high school students with and without learning disabilities. Twenty-one students (68% of the sample) had learning disabilities that affected their writing and 10 students did not have disabilities.

Two measures were used to evaluate accuracy of speech recognition: sentence probes and word-list probes. Students wrote essays under the following three conditions: using handwriting, using a scribe, and using speech recognition software.

The results indicate that two-thirds (68%) of the students achieved 85% accuracy and more than one-third (40%) achieved 90% accuracy using dictation to a scribe or speech recognition software. Only 3 students (10%) were below 80% accuracy. Results for adults have been reported between 90% and 98%. Results also demonstrate that both dictation conditions helped students with learning disabilities produce better essays. Students with learning disabilities produced higher quality essays when using a scribe, then when using speech recognition software. Both adapted conditions were better in quality than handwritten essays.