Prosodic and Linguistic Features of Child-Directed Speech in Word Learning Contexts: A Comparison Between Preterm and Full-Term Children

This database includes the raw data linked with the paper “Prosodic and Linguistic Features of Child-Directed Speech in Word Learning Contexts: A Comparison Between Preterm and Full-Term Children” under review on International Journal of Language & Communication Disorders. This study is part of a broader research program focused on early language acquisition and caregiver–child interaction in populations at neurodevelopmental risk. In this study, we investigate how maternal child-directed speech (CDS), in both its prosodic and structural components, varies between preterm (PT) and full-term (FT) toddlers and relates to early indicators of attention and vocabulary development. Procedures Forty-three mother–child dyads (20 PT, 23 FT) took part in the study when children were 18 months of corrected age (for PT) or chronological age (for FT). During a semi-structured play session, mothers were asked to label both familiar and unfamiliar objects while interacting naturally with their child. Maternal speech was audio-recorded, transcribed, and analyzed for structural features (verbosity, lexical diversity, syntactic complexity), while acoustic analyses (mean pitch, pitch range, speech rate) were conducted using PRAAT software on selected word tokens. Children’s visual exploration of the labeled objects was coded from video recordings, and receptive vocabulary was assessed using standardized tools. Analytical plan Group differences (PT vs. FT) in prosodic and structural CDS features were analyzed using independent-sample t-tests. Mixed-effects models were employed to test whether prosodic parameters varied as a function of word familiarity (familiar vs. unfamiliar referents). Correlational analyses were conducted within each group to explore associations between CDS characteristics, children’s visual attention, and vocabulary skills. Where relevant, regression models were used to adjust for potential covariates such as maternal education and infant sex. Findings in brief Mothers in both PT and FT groups systematically modulated prosody based on word familiarity, using higher pitch, wider pitch range, and slower speech rate when introducing unfamiliar objects. Compared to PT mothers, FT mothers used significantly more frequent and lexically diverse speech, though verbosity was similar between groups. In the FT group, slower maternal speech rate was associated with reduced child visual exploration, suggesting a sensitivity to children's attentional state. In contrast, no such association was observed in the PT group. However, in PT dyads, maternal verbosity, lexical diversity, and syntactic complexity were all positively correlated with children’s receptive vocabulary scores. These results point to shared prosodic strategies in word teaching across groups, but also to group-specific associations between maternal speech structure and language outcomes, particularly for children born preterm.