Turn-taking in free-play interactions: A cross-sectional study from 3 to 5 years.

Turn-transition timing in childhood has been examined by measuring response latency - that aggregates gap and overlap duration - in turn-transitions contingent to specific semantic categories. This contrasts with studies in infancy where the whole spectrum of temporal contingent vocalizations are examined, and gap and overlap duration is analyzed independently. We propose using the latter approach to investigate the continuities between infancy and childhood. In a cross-sectional design, we analyzed the vocalizations of 44 mother-child free-play interactions, ranging from three to five years of age. Frequency and duration were measured for gaps and overlaps, independently, and as an aggregated measure - floor-transfer offset (FTO). The effects of child's age and direction of turn-transition (child, mother) were assessed using generalized linear mixed modeling for each dependent variable (DV: FTO, gaps, overlaps). Although there was a slight increase in FTO and gap duration across ages, no significant effect of age was found for any of the DVs. There was an effect of turn-transition direction, for FTO and gap durations, but not for overlap duration. Specifically, mother-to-child turn-transitions produced significantly longer FTO and gap durations than child-to-mother turn-transitions, but had similarly timed overlaps. Results suggest that gaps and overlaps still have different developmental trajectories throughout childhood, and that overlap duration converges to adult standards, at least, by 3-years of age. Methodologically, we demonstrated the relevance of using complementary metrics (FTO, gap, overlap) to understand the developmental trajectories of turn-taking, and that examining all temporally contingent vocalizations can provide a valid and more inclusive measure of turn-transition duration in childhood.