Noakes, M. A., Schmitt, A. J., McCallum, E., & Schutte, K. (2019). Speech-to-text assistive technology for the written expression of students with traumatic brain injuries: A single case experimental study . School Psychology , 34 (6), 656–664. https://doi.org/10.1037/spq0000316

Journal Article

Noakes, M. A., Schmitt, A. J., McCallum, E., & Schutte, K. (2019). Speech-to-text assistive technology for the written expression of students with traumatic brain injuries: A single case experimental study. School Psychology, 34(6), 656–664. https://doi.org/10.1037/spq0000316

Tags

Dictated response; Dictated response (speech recognition system); Elementary; High school; K-12; Middle school; Traumatic brain injury (TBI); U.S. context; Writing

URL

https://www.apa.org/pubs/journals/spq/

Summary

Accommodation

Students' dictated response as documented using speech-to-text assistive technology (AT) was investigated for a written test-like task. These compositions were compared with students' handwritten products, responding to paper-based typed writing prompts. The voice recognition software was Dragon Naturally Speaking (DNS); students spoke into microphones connected to laptop computers after viewing writing prompts on computer screens. Researchers noted that participants completed the test-like tasks in separate quiet settings. 

Participants

Three (3) students with traumatic brain injuries (TBIs) of moderate to severe degrees participated. One student attended each of three grades—grades 4, 8, and 9—in a metropolitan area in a mid-Atlantic state (U.S.). Their ethnicity ("African American") and gender (male), and their current ages and ages at which they experienced brain injuries were also reported.

Dependent Variable

The narrative writing task was to respond to story starter writing prompts from the AIMSweb assessment system (Powell-Smith & Shinn, 2004). Participants' dictated compositions, and their handwritten products, were measured according to these written expression curriculum-based measurement criteria: total words written, words spelled correctly, and correct writing sequences. An alternating-treatment, single-case study design was used, with a series of 10 different writing prompts, across five weeks. The handwriting of participants' written compositions was described qualitatively, in terms of size and legibility, according to authors' observations in comparing their handwriting with their student peers.

Findings

Participants scored relatively low and unchanging across test sessions in the handwriting condition in total words written (TWW), with averages ranging from 2 to 23 among them. In contrast, with speech-to-text, average TWW scores ranged from 44 to 101 across participants, and larger between-session variation for each student. Comparisons yielded 100% non-overlapping data points and large effect sizes for all three students, favoring the assistive technology (AT) condition. In the handwriting sessions, students produced averages between 0 and 23 words spelled correctly (WSC) among them, with no variation to very low variation across test sessions. In contrast, with speech-to-text, students' WSC scores averaged between 44 and 101 among them, with much more—yet still relatively stable—variation across test sessions. Comparisons showed 100% non-overlapping data points and large effects sizes for all three students, favoring the AT condition. Finally, participants scored low and unchanging across the handwriting condition sessions in correct writing sequences (CWS), with averages ranging from 0 to 23 among them. In contrast, when using speech-to-text, students' CWS scores averaged 42 to 87, with larger variation. Again, comparisons indicated 100% non-overlapping data points and large effect sizes, favoring the AT condition. Participants' handwritten products were described as oversized and either unrecognizable letters and words, or under-developed lettering (in comparison with peers without TBI) with recognizable words. Limitations of the study were reported, and future research directions were suggested.