Exploring transcription processes when children with and without reading and writing difficulties produce written texts using speech-to-text

Abstract

Fluent transcription is hard to establish for children with reading and writing difficulties, due to problems with spelling. It has been proposed that composing by speech-to-text (STT) could facilitate their transcription, by circumventing the spelling process. To investigate this, transcription and error correction processes, and their relation to production rate (text length/time on task) was investigated in a sample of Swedish 10–13 year olds with and without reading and writing difficulties using STT to write expository texts. We determined effects of individual abilities: working memory, spelling, decoding, and the ability to interact with the STT tool under optimal conditions (STT success rate) on burst length, burst accuracy and production rate. Production rate was predicted by working memory capacity, by how long bursts the children produced, and by how accurate those bursts were. Further, burst accuracy was only predicted by a child’s STT success rate (in a test), but none of the other individual abilities. Dictating more than one word at a time and combining STT and keyboard use were identified as two useful strategies that can be taught in STT instruction. The results indicate that composing text using STT is a cognitively complex process placing heavy demands on working memory, and that STT success rate (that is, the combined effect of the technical capability of the STT tool and the participants output) is crucial to gain a fluent transcription without unnecessary disruptions.