Consistency and reliability of automated language measures across expressive language samples in autism.
Heather MacFarlane, Alexandra C Salem, Steven Bedrick and 8 others
PMID 36722653WHAT IT FOUND
Automated language scores from autism samples stayed stable over four weeks but changed significantly depending on whether the child was narrating, conversing, or answering ADOS questions.
Clinicians must compare scores only within the same task type, as a drop in complexity may reflect the setting rather than the child's ability.
Key findings
01Five of the six automated language measures showed significantly different distributions depending on whether the sample was taken during a conversation, a narration, or an ADOS interview.
02When the same task was repeated four weeks later, most measures showed good test-retest reliability, indicating that short-term fluctuations were minimal.
03Participants produced more complex language (higher Mean Length of Utterance) during the narration task compared to the conversation and ADOS tasks.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The sample included only individuals who could produce complex, fluent multiword utterances, so the results may not apply to minimally verbal or less fluent autistic clients. Most participants had IQs in the typical range, limiting generalizability to those with co-occurring intellectual disability. The study was exploratory with no multiple-comparison corrections, meaning some significant differences could be due to chance. Samples were collected in a clinical setting, which may not reflect natural conversational dynamics. The sample was heavily male (78.1%), reducing the power to detect sex differences.
Declared interests
The study was supported by the National Institutes of Health and non-U.S. government sources. No commercial conflicts of interest were reported in the provided text.
The easy way to misread this
Do not interpret lower automated language scores from an ADOS administration as evidence of a decline in a patient's language ability. The study shows that the ADOS context naturally elicits shorter utterances and more disfluency compared to narration or conversation tasks. A score drop may simply reflect the change in task demands, not a change in the patient's skill.