SES variation in language input and comprehension outcomes in Argentinean toddlers

ROSEMBERG, CELIA RENATA; ALAM FLORENCIA; STEIN, ALEJANDRA; MIGDALEK, MAIA JULIETA; MENTI, ALEJANDRA; SCAFF,C.; CRISTIA, ALEJANDRINA

Lyon

Congreso; IASCL; 2017

Previous research documents profound differences in both input and outcome as afunction of familial socio-economic status (SES; e.g., Huttenlocher et al., 2002). Mosthave studied relatively educated and rich populations, e.g. incomplete high schoolversus college graduates in the US. Current results regarding these educational ranges forSpanish learners suggest weaker relationships between SES and language acquisition(De Anda et al., 2015). Are effects of SES similar in other cultures and other educationalranges?We are carrying out a longitudinal project with 63 toddlers (mean age at the start: 14months), 36 from middle SES (mean length of parents? education: 20 years) and 27 from lowSES (mean length of parents? education: 9 years). We recorded their early linguisticexperiences repeatedly; 2 hours from an initial audio-recording gathered from 10 children ineach SES group have been fully transcribed. Additionally, vocabulary comprehension wasassessed about 15 months after the start of the study, with a touch-screen test built suchthat all stimuli had similar token frequency in middle and low SES households in Argentina,in order to prevent cultural bias affecting the children?s performance.Group comparisons suggest that low-SES children hear significantly lower numbers oftokens, less lexically diverse speech addressed in shorter utterances than mid-SESchildren; however, they overhear more speech, such that overall quantities are notsignificantly different. Moreover vocabulary comprehension scores are lower for low-SESthan mid-SES children. Ongoing analyses assess potential links between some aspects ofthe input and children?s comprehension performance.Overall, it appears that SES differences in language input and outcome reported in previousstudies are also evident in the present, widely diverse sample. Future work usingmeta-analytic methods may be able to assess whether SES differences are stable in sizeacross cultures and educational ranges.