Socio-economic Differences in Young Argentinean Children?s Language Environments

ROSEMBERG, C.; MENTI, A.; STEIN, A.; ALAM, F.; MIGDALEK, M.; OJEA, G.;FIBLA, L.; CRISTIA, A.

Nueva Orleans

Congreso; International Congress of Infant Studies; 2016

Over the last 40 years, many studies have described differences in quantity and quality in language input as a function of maternal education in North America (for a recent overview, see e.g., Hoff, 2013). This is a topic of both scientific and applied interest, since early input is thought to impact both early language development and later academic performance, at least according to American data (perhaps most famously, Hart & Risley, 1995). A great deal less is known about how language environment varies as a function of socio-economic status (SES) in other cultures. To contribute to filling this gap, we are gathering a large corpus from low and middle SES households in Argentina (N=30 in each). The former were recruited from extremely poor communities (?villas de emergencia?) and other very poor neighborhoods. These participants are mostly migrants from the North of Argentina or from neighbouring countries (Bolivia, Peru, Paraguay). Adults in these households had a low level of literacy, having completed 7 or fewer years of schooling. The middle SES group consists of families where one or both parents have some university education (>12 years of education). All the children are followed longitudinally over 18 months. The initial audio-recording, gathered with a small device worn by the child in a vest, is 4h long, of which the middle 2h were transcribed in CHAT format (MacWhinney & Snow, 1990).A total of 7 low SES (mean age 1;5.26; 3 boys) and 6 middle SES (mean age 1;6.20; 2 boys) children have been transcribed so far. The quantity and diversity of words heard by the child were analyzed using FREQ in CLAN (MacWhinney, 2015). The total number of word tokens in the child?s environment was significantly lower for the Low than the Mid SES children when only the mother?s speech was considered (p=.05), but significantly higher when summed separately over all other speakers (p=.01), with these trends cancelling each other out (Low mean 3850; Mid mean 3691; p=.78; see Figure 1). Similarly, lexical diversity did not differ across the groups (average across all speakers Low=.87; Mid=.89; p=.19).These results suggest that low SES households are not necessarily comparable across households. In our sample, children from low SES households suffered from neither lower overall levels of verbal activity nor lower lexical diversity than their middle-SES peers. Further inspection of our transcripts suggested that children growing up in the low SES households we investigated did not suffer from the social isolation described among American low SES environments e.g. in Hart and Risley (1995). Quite to the contrary, our low SES children came in contact with twice as many children and adults than their middle SES peers. This leaves many questions open, including whether Argentinean low SES children also benefit from other behaviors that have been described as boosting language acquisition (such as contingency) to a similar extent as mid SES Argentinean and American children, which we are addressing in follow-up analyses.