How high school students used speech-to-text as a composition tool

Abstract

Speech to text (STT) technology, also called voice recognition or speech recognition technology, automatically transcribes users’ speech to a computer screen. Research indicates that STT shows promise as an alternative mode of composition, supporting students in making fewer errors, writing more, and writing better. Much of this research takes place in elementary and middle schools, often focusing on students with identified learning disabilities. The current study extends STT research into general education high school classrooms, where students (N = 120) were invited to use it as much or as little as they wanted during one school year. We asked: Who used STT? Did subgroups, such as students labeled with learning disabilities, differ in their choices? When students chose to use STT, what kinds of writing did they do? Did STT compositions differ from non-STT compositions? Finally, what were students’ and teachers’ perceptions of STT? We analyzed field notes, interviews, surveys, and student writing with and without STT, and found mixed results: On average, students wrote more when they used STT than not. A little more than half of students found that STT eased the cognitive load of composition. A little less than half of students found that the tool constrained their composition. Some were put off by technical problems or embarrassment attendant with speaking out loud in class. Students’ choices to use STT correlated with special education designation, but not other designations. Despite students’ mixed reports, teachers were consistently positive about STT and planned to use it in future classes.