Predicting Meaningful Differences in School-Entry Language Skills from Child and Family Factors Measured at 12 months of Age

Abstract

Early childhood services which seek to promote early language development are hampered by the absence of reliable methods to identify children who may develop persistent language difficulties. This is because of variability in preschool children’s language development and that existing measures have limited diagnostic accuracy. In this study, we examined two approaches to identifying children at 12 months of age who are at risk of poor language skills at 4 years of age: (1) measurement of socio-cognitive and communicative foundations of language using the “Sure Steps to Talking” (SSTT) Questionnaire; (2) a “combined risk model” utilising child, family, and parenting characteristics. The study had three phases, and they are as follows: (1) evaluation of the predictive validity of SSTT through longitudinal follow-up of children in a matched cohort in the UK; (2) modelling application of SSTT as a population screener and assessing its validity in a longitudinal community cohort in Australia (Early Language in Victoria Study ELVS); and (3) evaluation of a list of risk factors through secondary data analysis of the ELVS cohort. Prediction using the SSTT measure was modest. A “risk model” that included seven factors (child communicative behaviours, family, and parenting characteristics) measured at 12 months reached “fair” levels of predictive validity and approached a similar level of prediction for language difficulties at age 4 years as found by measures of vocabulary at 24 months. This latter assessment approach potentially provides an additional 12-month window within which preventative interventions could be implemented.