Speech-to-Text Technology¶

Strategy · All strategies
Evidence · 1 claim (1 for) · 2 studies (1 causal, 1 quant-synthesis), q3–q4 · 1 of 2 report an effect size

Description¶

Speech-to-text (STT) technology converts spoken language into written text in real time. The learner speaks into a device (computer, tablet, or smartphone) equipped with speech recognition software, which transcribes the words; the learner then corrects errors and adds punctuation. For learners with dyslexia, dysgraphia, or physical disabilities, STT removes the transcription barrier — spelling, letter formation, and typing mechanics — so composing effort can go to ideas and organization rather than transcription.

Design Implications¶

STT functions as an accommodation that bypasses transcription demands, freeing working memory for planning and generating content Cognitive overload degrades learning when transcription demands compete with composition. [+M]. Its effectiveness depends on the learner treating dictation as a distinct composing skill: speech is more oral and less structured than writing, so instruction must cover planning before speaking, explicit punctuation commands, and systematic proofreading of the transcript. Dictation without instruction in these subskills often produces rambling, error-laden text that does not improve writing quality.

Context¶

Requirements¶

• A device with a microphone and speech recognition software (built-in dictation, Google Docs Voice Typing, or dedicated software such as Dragon)
• A quiet environment; background noise degrades recognition accuracy
• Initial training: voice enrollment where available, learning verbal punctuation and editing commands, and practice speaking in composed phrases rather than fragmented words
• An explicit proofreading routine, since recognition errors are plausible-looking and easy to miss

Constraints¶

• Accuracy drops with background noise, accents, atypical speech, and complex or technical vocabulary [-M] — errors can exceed the learner's ability to detect them, especially for learners with weak spelling knowledge
• Composing by voice changes the register of the text; untrained dictation tends toward run-ons and oral phrasing, so quality gains are not automatic [~M]
• Editing the transcript requires reading skills; for learners whose reading is as impaired as their writing, STT can shift rather than remove the barrier
• Over-reliance without gradual integration of spelling and typing practice may leave transcription skills underdeveloped [~W]

Implementation Variability¶

• Compensatory use: STT as a standing accommodation for all written work, prioritizing access to grade-level composition tasks
• Remedial-integrated use: STT paired with explicit writing instruction, with dictation used for drafting and conventional writing practiced for shorter tasks
• Universal access: STT built into mainstream tools (Google Docs Voice Typing, Apple Dictation, Windows dictation) so all students can use it without stigma, consistent with Universal Design for Learning approaches
• Low-stakes first: introducing STT for brainstorming and note-taking before high-stakes assignments

Target Learners¶

• Students with dyslexia or dysgraphia whose transcription (spelling, handwriting) is markedly weaker than their oral language [+M]
• Students with physical disabilities affecting handwriting or typing [+M]
• English learners and students with speech differences may benefit less or need voice training, as recognition accuracy is lower for accented or atypical speech [~M]
• Less necessary for learners with fluent transcription skills, for whom dictation offers little advantage and may interfere with revision habits [~W]

Target Learning Goals¶

• Written expression: producing longer, more complex compositions than the learner can produce by hand [+M]
• Composition fluency: increasing output rate so drafting keeps pace with idea generation
• Content-area access: completing essays, responses, and research writing in subjects where transcription was previously the bottleneck

Instructions¶

1. Assess whether transcription (not ideation) is the actual barrier; STT helps most when oral language exceeds written output
2. Teach the tool explicitly using Direct Instruction: voice commands, punctuation, correcting errors, and speaking in full composed sentences
3. Model dictation aloud, thinking through planning before speaking (Think-Aloud), so learners hear what composed dictation sounds like
4. Have learners practice with low-stakes tasks — brainstorming, journal entries, short responses — before graded writing
5. Establish a proofreading and editing routine with a checklist; provide feedback on both content and dictation technique
6. Review progress periodically and decide whether to maintain, fade, or expand STT use

Related Strategies¶

• Word Prediction Software — the complementary transcription support for learners who type but spell poorly
• Text-to-Speech Technology — the inverse accommodation; pairing the two lets learners hear and check their dictated text
• Graphic Organizers — pre-dictation planning structures that compensate for the linear nature of speech

Examples¶

• Dragon Speech Recognition (Nuance) — professional-grade dictation with voice training and custom vocabulary; widely used for students with severe dysgraphia or motor impairments
• Google Docs Voice Typing (https://support.google.com/docs/answer/4492226) — free, built-in dictation that makes STT available to whole classes without stigma
• Apple Dictation and Windows Speech Recognition — built-in OS-level dictation across devices
• A middle-school student with dyslexia dictates first drafts of essays via Voice Typing, then edits with text-to-speech readback — dictation for generation, readback for revision

Key Sources¶

• De La Paz, S., & Graham, S. (1997). Effects of dictation and advanced planning instruction on the composing of students with writing and learning problems. Journal of Educational Psychology, 89(2), 203–222.
• MacArthur, C. A. (2009). Reflections on research on writing and technology for struggling writers. Learning Disabilities Research & Practice, 24(2), 93–103. doi:10.1111/j.1540-5826.2009.00283.x
• Graham, S., & Perin, D. (2007). A meta-analysis of writing instruction for adolescent students. Journal of Educational Psychology, 99(3), 445–476. doi:10.1037/0022-0663.99.3.445
• Evmenova, A. S. (2018). Preparing teachers to use universal design for learning to support diverse learners. Journal of Digital Learning in Teacher Education, 34(4), 227–244. doi:10.70725/806763ugjaeu