Cross-Situational Statistical Word Learning in Late Language Emergence: An Online Study

Document Type

Peer-Reviewed Article

Publication Date

2025

Abstract

Purpose: Cross-situational statistical learning is one mechanism by which typically developing toddlers map words to referents. Yet, this type of statistical learning has been found less efficient in children with developmental language disorder (DLD). The purpose of this article is to evaluate cross-situational statistical learning in very young children with language delay, late talkers (LTs), compared to typically talking toddlers. We predict that LTs will show inefficiency in cross-situational statistical word learning similar to older children with DLD.

Method: LT (n = 15, 18-34 months) and typical talker (TT; n = 15, 18-35 months) groups matched on chronological age and sex completed a cross-situational statistical learning task in which they were trained on six novel word-referent pairs and then tested on these word-referent associations. The experiment was completed on the participant's home computer, and gaze was recorded for the duration of the experiment. Mixed-effects models were used to evaluate group differences in time spent looking at labeled referents as a measure of learning.

Results: The LT group spent an equal proportion of time looking at the named targets and the unnamed distractors when tested, suggesting minimal learning had occurred. The TT group, in contrast, spent a significantly greater proportion of time looking at the targets when labeled, indicating more established word-referent links.

Conclusions: These findings suggest that LTs, like older children with DLD, are less efficient at leveraging cross-situational statistical learning opportunities that may, in addition to other factors, contribute to their slow expressive vocabulary development.

Comments

Online ahead of print, March 28, 2025

At the time of publication, Olivia Cayward was a graduate student in the College of Health Professions

DOI

10.1044/2025_JSLHR-24-00670

PMID

40152669

Link to Full Text


Share

COinS