Digital Promise’s Dr. Xin Wei recently analyzed the National Assessment of Educational Progress Grade 8 mathematics data to examine the TTS behaviors of more than 2,000 students who chose to use the tool on a math item. Rather than asking only whether students used TTS, the study looked more closely at three dimensions of use: how long students listened, how often they activated the tool, and the sequence in which they used it.
The findings point to a simple but important lesson for educators and designers: TTS is most useful when it is used purposefully, and its value differs across students.
Students who used TTS were more likely to be lower-performing, to have disabilities, to be English learners, and to receive extended time accommodations. In other words, the students who turned to TTS were often those who may have needed additional support accessing the text of the math item.
But the most interesting result is that TTS was not equally helpful for everyone.
For students performing at the Below Basic level, spending time listening to the problem statement was associated with a higher chance of answering correctly. The relationship was not unlimited, however. Accuracy improved as listening time increased, but only up to about 25 seconds; after that, the benefit leveled off. This suggests that TTS can help struggling students gain access to the meaning of a problem, but more listening is not always better.
For Basic-level students, a different pattern emerged. Students who repeatedly turned TTS on and off without sustained listening were less likely to answer correctly. This kind of toggling may reflect distraction, uncertainty, or inefficient strategy use. In practical terms, it suggests that TTS can become less helpful when students use it in a fragmented way.
For students at the Proficient or Advanced levels, there was no consistent relationship between TTS use and accuracy. These students may already be able to process the text efficiently on their own, so TTS may add little value in this context.
The first implication is that schools should move beyond a simple “access provided” mindset. Offering TTS is important, but it is only the first step. Students also need support in learning when and how to use it effectively.
Second, teachers and support staff may want to explicitly model strategic TTS use for struggling readers. For example, students can be taught to listen carefully to the problem statement when a math item is linguistically dense, rather than repeatedly clicking in and out of the tool.
Third, practitioners should not assume that the same support works the same way for every student. TTS appears to be especially useful for some lower-performing students, but less relevant for stronger readers. This is exactly the kind of variability that Universal Design for Learning asks us to take seriously: learners benefit from options, but not all options are equally effective for all learners.
Assessment and platform designers should pay attention to interface design. If repeated toggling is linked to poorer performance for some students, platforms could be designed to reduce nonproductive use—for example, through clearer controls, smoother playback, or prompts that encourage more sustained listening.
These improvements are likely to be most effective when they are developed through co-design with students and educators. Co-design helps ensure that accessibility features are not only available, but also intuitive, supportive, and responsive to how learners actually use them in practice.
The broader message is straightforward: assistive technology works best when access is paired with guidance, practice, and thoughtful design. TTS is not a magic fix, but it can be a meaningful support, especially for students who need help accessing the language of mathematics.
To learn more about the study and its implications for accessibility and assessment design, read the full study here.
