SLPOtherJournal of speech, language, and hearing research : JSLHR2026

Does the Use of Crowdsourced Listeners Yield Different Speech Intelligibility Results Than In-Person Listeners for Typically Developing Children?

Heather D Salvo, Tristan J Mahr, Carly Sandgren and 2 others

PMID 41610401

WHAT IT FOUND

Crowdsourced listeners scored child speech as less intelligible than in-person listeners, with the largest drop for children in the mid-range of clarity.

This method underestimates performance compared to standard norms, so it should not replace clinical assessment for these patients.

Key findings

01Crowdsourced listeners produced intelligibility scores up to 7 percentage points lower than in-person listeners, even with strict quality checks.

02The discrepancy between listener types was greatest for children with mid-range intelligibility (65% to 83%).

03Using crowdsourced scores against in-person norms caused most children to drop in age percentile rank, potentially misclassifying them as impaired.

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.

Already have one?

What it does not show

The study used typically developing children, so the results may not generalize to patients with dysarthria or other speech disorders where error patterns are less predictable. Crowdsourced listeners did not undergo technical headphone checks, relying only on self-report, which may have introduced uncontrolled audio quality variations. The in-person listeners were a homogeneous group of university students, while crowdsourced listeners were more diverse in age and background, which may have influenced transcription accuracy independently of the listening environment.

Declared interests

The study was funded by the National Institute on Deafness and Other Communication Disorders and the National Institute of Child Health and Human Development. The authors declared no conflicts of interest.

The easy way to misread this

Do not assume that crowdsourced data is equivalent to in-person data simply because the trends look similar. The absolute scores are lower, and applying these scores to existing clinical norms will falsely classify many children as having intelligibility disorders.

Summarised by AI from the full paper, without a clinician reviewing it. Check it against the source before it changes what you do. Read it on PubMed →