Concurrent and predictive validity of parent-report measures of children's language development

Abstract

Introduction. The study aimed to inform the question of how professionals should respond to parents who approach them with concern about their child’s speech and language development, particularly after they have been primed to think about their child’s language. Methods. Data from a longitudinal investigation of 168 English-speaking children are reported. Cohorts of children with (n = 51) and without (n = 117) early language delay were assessed at four time points in their development between the ages of 2 and 5. Parents completed three written questionnaires at Time 1 when the children were 24-31 months of age: 1) a New Zealand adaptation of the MacArthur-Bates Communicative Development Inventory: Words & Sentences; 2) the Parent’s Evaluation of Developmental Status; and 3) a parent questionnaire. Children were also administered the Preschool Language Scale 4 (PLS4), a standardized language test, at Time 1 and Time 3, 18 months later. Results. The associations between various combinations of parent-report measures and clinical outcomes were estimated with likelihood ratios. At Time 1, children who were not combining words and whose parents expressed concern were 38 times more likely to have PLS4 Expressive Communication scores in the low range (1 SD or more below the normative mean) than in the average range. By Time 3, there was still a 9-fold increase. The positive and negative likelihood ratios of all parent-report measures at each time point will be reported. Conclusions. We conclude by highlighting challenges faced in translating research evidence into clinical practice and suggest a next step in the research.