Language Samples From Children Who Use Speech-Generating Devices: Making Sense of Small Samples and Utterance Length.

Area of Science:

• Speech-language pathology
• Linguistics
• Assistive technology

Background:

• Mean length of utterance in morphemes (MLUm) is crucial for assessing language development but is underreported in individuals using augmentative and alternative communication (AAC).
• Key challenges include obtaining representative language samples and accurately transcribing very short utterances common in AAC users.
• Speech-generating devices are frequently used by children requiring AAC interventions.

Purpose of the Study:

• To address the underreporting of MLUm in AAC users by testing solutions for common measurement challenges.
• To evaluate the effectiveness of adjusted sampling windows for obtaining representative language data.
• To assess the utility of mean syntactic length (MSL) as an alternative metric for utterance length.

Main Methods:

• Language samples from children using speech-generating devices were analyzed.
• Sampling window duration was adjusted to overcome the challenge of small language samples.
• Mean syntactic length (MSL) was employed as a proxy for MLUm to address transcription difficulties.

Main Results:

• A 24-hour sampling window was insufficient for representative samples, while a 1-month window consistently yielded representative data.
• Mean syntactic length (MSL) significantly and positively predicted MLUm in a normative sample.
• MSL measures from representative samples accurately predicted MLUm in children who use AAC.

Conclusions:

• Valid and reliable measures of utterance length for AAC users can be achieved through extended sampling periods (1 month) and the use of MSL.
• Further research is necessary to fully understand the strengths and limitations of these proposed solutions for AAC language assessment.