Unleashing the potential in young writers: Using speech recognition software with students with writing difficulties
Research suggests that difficulties in low-level transcriptional processes such as handwriting and spelling or capitalisation can affect the quality and fluency of text and restrict children’s capacity to generate and record ideas (Graham and Perin, 2007). Using speech recognition software (SRS) provides an opportunity for students to dictate, edit and review their work using voice commands as well as typing.
SRS may therefore enable students who find handwriting a barrier to their learning, including those with writing difficulties and other special educational needs and disabilities (SEND), to circumvent transcriptional difficulties such as poor spelling and motor skills by converting speech into text, and, as a result, may improve the quality and quantity of students’ writing. In doing so, it raises their motivation to write better (MacArthur , 1996), faster, with less fatigue than typing and with a varying degree of independence (Cochran-Smith, 1991). Moreover, previous meta-analysisA quantitative study design used to systematically assess the results of multiple studies in order to draw conclusions about that body of research of true and quasi-experiments demonstrated that writing instruction improves students’ reading skills, and that writing about material read or presented in class enhances the learning of such information (Bangert-Drowns et al., 2004); (Graham and Hebert , 2011).
In their review of interventions to improve the written-language performance of children with Asperger Syndrome, Schneider et al. (Schneider et al., 2013) suggested that future research should focus on exploring SRS, and indeed research in the applications of SRS is still accruing. Despite significant difficulties thus far with the quality of the technology in capturing nuances in the voice, as SRS technology is rapidly improving, so will the opportunities to use it more efficiently: SRS is now available to students in further education and, since 2015, the technology has also been used in GCSE exams.
Our research
During the last few years, we have carried out a series of classroom-based interventions, following a mixed-method single-case design, which aimed to investigate the potential impact of SRS use on students’ writing. The interventions targeted specific students who, according to their teachers, found writing difficult and who would benefit from a new approach to learning the skill. MA teacher-researchers followed closely and recorded these interventions and collaborated with the class teachers and TAs in their everyday tasks (Simon , 2014); (Little , 2017).
A total of 28 students aged between eight and 16, with varying levels of writing skills difficulties and other SEND, were selected by their teachers and TAs in one special and two mainstream schools. First the teachers and then the students were trained to use SRS (the intervention), and the changes in their academic performance were measured over a period of between six and nine weeks. The purpose of using the SRS was explained and modelled to the students, and they then set up a voice file by reading from a set passage. They finally practised using voice commands, such as ‘full stop’ for punctuation, or ‘new line’ or ‘go to end’ for navigation in the text. Each child read passages from their reading books to practise (four sessions of 30 minutes each), though training varied in depth according to the child’s needs. Two tasks were given pre- and post-intervention: a handwriting task and then an SRS task (spoken and edited text).
Progress was measured pre- (baseline) and post-intervention, using TOWL-4 (Hammill and Larsen , 2009) – a validated test of written language based on three writing components (conventional, linguistic and conceptual) and two writing assessment formats (contrived and spontaneous). The subtests assess vocabulary, spelling, punctuation, logical sentences, sentence combining, contextual conventions (writing a story in response to a picture) and story composition. Qualitative data was gathered through interviews with teachers, and student observations were carried out in the beginning and at the end of the study to supplement the picture.
Preliminary findings
We are still in the process of aggregating the test results for writing, but preliminary results by school seem positive, with improvement in the students’ writing skills as measured by the TOWL. The evidence of improvement comes also from the qualitative data provided by the teachers. TransferThe processes of applying learning to new situations of skills such as punctuation were noted, particularly in the special school, where the commands used for the SRS were repeated as rhymes by the children and applied without help.
Growing technology-confident teachers
The role of the teacher and teaching assistants, plus effective communication and coordination between staff, emerged as important factors that could contribute to the successful long-term use of SRS. The findings point to the feasibility of SRS to support children with writing difficulties when used with strong instruction, and its most effective use therefore to be when placed firmly in the hands of confident pedagogues.
