Inequalities in Vocabulary: Examining the Predictive Value of SES Measures Over Developmental and Historical Time.

• In Event: Conditions of social vulnerability and language development

Abstract

It has long been known that there is a relationship between socioeconomic status (SES) and language development (e.g., Clegg & Ginsborg, 2006; Hart & Risley, 1995; Labov, 1969). A recent resurgence in interest in this phenomenon has focused on the so-called word-gap, whereby those from higher SES backgrounds tend to have a larger vocabulary (e.g., Fernald, Marchman, & Weisleder, 2013; McGillion, Pine, Herbert, & Matthews, 2017; Pace, Luo, Hirsh-Pasek, & Golinkoff, 2017). Interventions to close the gap have largely focused on parenting interventions to support very young children. Results have been mixed and we still do not know how to best address inequalities, with major studies run decades apart in a shifting economic and social landscape, and little empirical evidence examining whether such interventions have had any effect. To make better targeted progress, arguably we first need to better understand the nature of the relationship between SES and language. In a step towards this goal, we looked at this relationship in two nationally representative cohorts with pre-registered analyses that addressed three fundamental questions:
1) Which measures of SES are most closely related to child language outcomes?
2) Are associations stable across developmental time, or does the gap widen or narrow with age?
3) Are associations stable across historical time, or has the gap widened or narrowed since the 1970s?

We used data from the UK Millennium Cohort Study (a longitudinal birth cohort study of 19,518 young people) to investigate the value (both in isolation and in a composite model) of five separate indicators of SES: parent education, income, wealth, occupational status and relative neighbourhood deprivation in predicting vocabulary. Again using Millennium Cohort Study data, we addressed how the relationship between each SES indicator and language changed over developmental time, looking at the relationships at ages 3, 5, 11 and 14. Finally, by comparing the Millennium Cohort to the UK 1970 British Cohort Study (a longitudinal birth cohort study of 16,571 individuals born during one week in 1970), we investigated whether there have been changes in the relationship between SES and language for cohort members born 30 years apart.

Findings revealed parent education as the SES predictor that explained the most variance in the language outcome. However, income and occupational status also predicted substantial unique variance in the language outcome at all ages. A composite measure explained more variance in language at each age compared to any one predictor of SES considered in isolation, but considering multiple SES indicators simultaneously explained more variance than the composite measure. Regarding developmental change, we observed that inequalities in language ability are evident from the age of 3 and widen between the ages of 3 and 5 (see figure 1). Inequalities then remain persistently wide throughout childhood and into adolescence. Finally, regarding historical change, our cross-cohort comparison suggests the social gradient has been roughly stable, although inequalities based on occupational status were widest in the older 1970 born cohort and the advantage of university level education is clearest in the 2001 cohort (see figure 2).

Authors

• Emma Thornton, University of Liverpool
  Presenting Author

• Colin J. Bannard, The University of Liverpool
  Non-Presenting Author

• Danielle Elizabeth Matthews, University of Sheffield
  Non-Presenting Author

• Praveetha Patalay, University College London
  Non-Presenting Author