MABEL Phoneme Awareness Tests
The Psychometrics Centre has collaborated with speech and language researchers at Bangor University to develop computer-assisted assessments of phoneme awareness skills for the MABEL web-based assessment tool (eldel-mabel.net), which can be administered in several different languages. Using the Concerto platform, the Centre created online versions of two gold standard tasks, namely an analytic phoneme awareness skill, called Phoneme Deletion, and a synthetic phoneme awareness skill, called Phoneme Blending.
The assessments work by allowing an administrator to playback audio recordings of stimulus words to the participant, with a degree of automation, and then recording the participant’s response using the device microphone. Both assessments capture response latency at the item and task levels, which is used for scoring, and provide a convenient means for administrators to download the audio files and later score the participant’s spoken responses.
The reliability of phoneme awareness measures is strongly dependent on the accuracy with which the administrator pronounces the stimuli, and the level of ambient noise during the test. By allowing administrators to play pre-recorded stimuli to participants rather than speaking words themselves, the technical innovations brought by the Psychometric Centre’s computer assisted test versions bring many advantages, including a stable, high quality presentation that can be carried out via headphones. Automatic timing of responses, and auto-skipping to the next item also allows for greater consistency in administration time and reduces the burden on administrators, who would otherwise need to run the test with a stopwatch and instruction manual to hand.
MABEL is a tool intended for use in many different language contexts. Therefore, these computer-assisted formats bring a huge benefit of international reach to literacy practitioners and researchers who may carry out the assessments with high fidelity in the language of their choosing.
Future collaboration on this project is underway to explore the development of further language versions as well as the possibility of transcribing (in addition to recording) the participant’s responses using a speech language transcription service, with the aim of aiding or even automating the comparison between spoken responses and expected responses.