Frequency effects on screening accuracy (Poll et al., 2025)
Purpose: Sentence repetition (SR) is a promising task for identifying children at risk for developmental language disorder (DLD) but is unclear how to calibrate the task for adults. In verbal recall, the frequency of language structures affects adults with DLD differently from peers with typical language (TL), particularly near the limits of adults’ processing capacity. We hypothesized that repetition accuracy for sentences with less frequent structures, passives, would better distinguish adults at risk for DLD from their peers with TL than sentences with more frequent structures, actives.
Method: Forty-two adults (18–29 years old), including 17 classified as at risk for DLD, completed an SR task composed of active and passive sentences matched for lexical frequency and sentence plausibility. Sentences varied in length (eight to 16 words) and number of clauses (one to two). Repetition accuracy for matched active and passive sentences was evaluated for classification accuracy to determine whether the areas under the curve (AUCs) differed by structure.
Results: Averaging repetition accuracy across all sentence lengths, the AUC for passives did not differ from actives. For sentences with 11 words and two clauses, the AUC for passive structures was significantly higher than for active structures.
Conclusions: Given sentences long enough to challenge the capacity of adults to reconstruct sentences for recall, structural frequency affects those at risk for DLD differently than those with TL. Manipulating the frequency of sentence structures is a promising approach to developing an SR task suited to screening adults.
Supplemental Material S1. Data and statistical comparisons of lexical frequency, lexical age of acquisition, and sentence plausibility across sentence conditions.
Poll, G. H., Brown, B., & Miller, C. A. (2025). Exploring the effect of sentence structure frequency on the accuracy of a screener for adults at risk of developmental language disorder. Journal of Speech, Language, and Hearing Research, 68(7), 3348–3356. https://doi.org/10.1044/2025_JSLHR-24-00628
