Building a typical and disordered speech dataset for teaching and assessing phonetic transcription in speech and language therapy students in Ireland.

Abstract

Corpora of speech sound disorder are precious resources in Speech and Language Therapy (SLT) education, but their development is costly in terms of time and funding. This paper reports on an initiative that created a dataset of Irish English-based materials for teaching and assessing phonetic transcription at the pre-registration level. The dataset includes audio recordings of single words, sentences and spontaneous speech elicited from five children with speech sound disorder, five children with typical speech development and five typical adults with no speech difficulties. In order to design and evaluate the resource, SLT students and Practice Educators were surveyed before and after resource creation. We report on the overall process, discuss results from the surveys and outline future directions in resource development for teaching purposes. Overall, our respondents agreed that the resource filled a gap in teaching materials for SLT education.