Nelson, L. M., & Reynolds, T. W. (2015). Speech recognition, disability, and college composition . Journal of Postsecondary Education and Disability , 28 (2), 181–197. https://www.ahead.org/publications/jped
Nelson, L. M., & Reynolds, T. W. (2015). Speech recognition, disability, and college composition. Journal of Postsecondary Education and Disability, 28(2), 181–197. https://www.ahead.org/publications/jped
[no doi reported]; Also downloadable from ERIC online database: https://eric.ed.gov/?id=EJ1074665
Perceptions about the dictated response accommodation, in the form of speech recognition assistive technology (provided through Dragon Naturally Speaking software, version 11), were investigated.
Participants were five postsecondary students with disabilities who were registered with the Disabilities Support Office (DSO) at the community college in Virginia. Their disabilities included: Attention Deficit Hyperactivity Disorder (ADHD), ADHD and mental health disability, cerebral palsy and ADHD, dyslexia, and learning disability related to language. Other demographic and personal characteristics were reported, including gender, race, and speech recognition experience.
The task scoring rubric for evaluating the composition work product included overall quality as well as spelling, use of vocabulary, and errors. Researchers also documented their observations of participants while doing their work. Interview questions included background experiences of students with writing and with using supports in general and speech recognition software in particular. Participants also reflected on their use of this accommodation while writing their compositions.
Postsecondary student participants reported varying amounts of experience using speech recognition tools prior to the current writing task. Beginning users seemed to adopt the software use "very quickly and with little difficulty" (p. 189). Some challenges included clear enunciation, even after the software was trained to users' voices, and becoming accustomed to continuing to speak without being distracted by viewing the screen even when the transcriptions had incorrect grammar or mistakes in word choice. They determined that use of the keyboard and mouse at times became necessary when editing, rather than using voice commands only. These new users remarked that using this support made composing quicker and easier than typing (manually), and reduced the likelihood of their becoming tired early in the task. For at least one new speech recognition user, the student with ADHD and mental health disabilities, the software allowed there to be better spelling and fewer errors that typically have slowed down her writing process. Two participants indicated substantial experience with speech recognition in middle school and high school, and one indicated using a sort of conversational approach, considering the computer a human listener. Both of these students noted their having learned not to stop after each sentence to make spelling and grammar corrections, so as not to disrupt their trains of thought. They also both became practiced at organizing their thoughts without pre-planning or using written outlines or notes. They both indicated that they performed editing by using the keyboard rather than by voicing edits to their computers. Limitations of the study were reported, and future research directions were suggested.