The “landscape of talk” in home language environments of Spanish-speaking families with young children
Publication Date
7-3-2018
Document Type
Presentation
Publication Title
International Congress of Infant Studies Biennial Congress
Conference Location
Philadelphia, PA, United States
Abstract
How much caregivers talk to their children has profound implications for children's language learning. Children from lower-socioeconomic (SES) families tend to hear less child-directed speech (CDS) than children from higher-SES backgrounds, which is associated with poorer language and academic outcomes (Hart & Risley, 1995). Many studies have observed caregiver-child interactions during contexts such as play or book reading in a laboratory setting, and some have examined interactions at home during dressing routines or mealtimes (Hoff-Ginsberg, 1991). Recently, Tamis Le-Monda et al. (2017) examined in-home spontaneous contexts over 45 min in middle-high-SES English-speaking families. Yet, we know little about how often such contexts are represented spontaneously over a typical day, especially in lower-SES, non-English-speaking families. We recorded caregiver-child interactions in Spanish-speaking families from diverse SES backgrounds, capturing spontaneous periods of dense talk using the LENA technology. Taking 3 hrs of each recording with the most adult talk, we examined the frequency distribution per child of five primarily-CDS contexts (Unstructured Interactions, Meals, Books, Play, Routines), asking whether amount of talk differed across contexts, and whether talk was related to child language outcomes. Participants were 87 lower-SES Spanish-speaking families with 25-month-old children. Native Spanish-speaking coders listened to the 3 highest-density hrs of each day-long LENA recording, focusing on those segments with the highest adult word counts (AWC), and classified 18 10-min interactions into one of 5 CDS contexts. Children's online understanding of familiar Spanish words (Accuracy and RT) was assessed with the "looking-while-listening" (LWL) procedure, and parents reported their child's vocabulary size (Spanish MB-CDI). Figure 1 shows the mean proportion of different contexts per child (line) and mean AWC counts per context (bar). The most frequently occurring context was Unstructured Interactions, while Books and Routines occurred infrequently. In fact, fewer than half the children had any Books contexts at all. Mean AWC was significantly higher in Books than all other contexts (ts>6.48, ps<.001), with no differences among the other contexts (ts<1.13, ps>.79). Intercorrelations in AWCs across contexts were significant (rs=.45-.87), indicating that children who heard more talk in one context also heard more talk in other contexts. Finally, those children who heard more talk overall were faster (r= -.37, p<.001) and more accurate (r=.33, p=.002) in the LWL task. Average amount of talk was uncorrelated with reported vocabulary size (r=.11). Ongoing analyses explore patterns of talk across the different contexts, whether AWC differs within context type as a function of whether caregivers and children are jointly interacting, and the quality of caregiver talk. This is a first look at the "landscapes of talk" in spontaneous, in-home interactions between Spanish-speaking parents and their 2-yr-old children. Periods of dense talk frequently occurred in unstructured interactions, rather than in traditional laboratory contexts. Children heard the most caregiver talk during book reading, which was relatively infrequent. These results provide insights into when and how Spanish-speaking caregivers spontaneously talk to their children and can inform ways to support rich caregiver-child engagement in families from diverse backgrounds.
Department
Child and Adolescent Development
Recommended Citation
Janet Bang, Monica Munévar, J. Douglas Mastin, Virginia Marchman, and Anne Fernald. "The “landscape of talk” in home language environments of Spanish-speaking families with young children" International Congress of Infant Studies Biennial Congress (2018).
