Digital Promise tested how text-to-speech features behave in standardized assessments and found a nuanced impact that varies by student subgroup. In an analysis of 8th graders’ use of read-aloud functionality on a geometry question from the National Assessment of Educational Progress, only 7% of students (about 2,000 of roughly 26,000) used the feature. Students who used text-to-speech were more likely to be English learners, students in special education, students of color, and students receiving extended testing-time accommodations. The study reported that lower-performing students were more likely to get the question correct when they listened longer, with an effect plateauing at about 25 seconds. The findings suggest a double-edged role—helpful for some students when used appropriately, but potentially distracting for others who toggle the tool repeatedly. Digital Promise researcher Xin Wei said educators should treat text-to-speech like a problem-solving strategy and build proficiency ahead of test day.