Validity and Reliability of Mothers' Reports of Language Development in 1-Year-Old Children in a Large-Scale Survey in Bangladesh

Abstract

BACKGROUND: In developing countries, it is often important to have measures of development in children under 3 years of age in large-scale surveys or evaluations of nutrition and stimulation programs. However, there is a lack of suitable instruments with established validity. OBJECTIVE: To develop a language test for children aged 12 to 18 months based on mothers' report, suitable for use in large-scale surveys, and examine its concurrent and predictive validity. To determine whether the test is sensitive to home stimulation and nutritional status and compare the test with the Bayley Scales of Infant Development (BSID). METHODS: A subsample of participants in a large, prospective cohort study in rural Bangladesh (MINIMat) was selected for a study of child development (n = 2,852). A total of 2,418 participants were interviewed concerning their children's expressive and receptive vocabulary, and children were tested using the BSID. RESULTS: The language test had reasonable short- and long-term reliability between 12 and 18 months (r = 0.50) and concurrent validity with the Bayley Mental Development Index (MDI) (r = 0.32 language comprehension to 0.41 language expression). Its predictive validity with IQ at age 5 years was similar to that of the Bayley MDI (r = 0.37 to 0.41 for language and r = 0.37 for MDI). Child language was independently associated with postnatal growth, stimulation in the home, gestational age, and socioeconomic status, and a similar set of variables predicted the Bayley MDI. CONCLUSIONS: The language test was reliable, had acceptable concurrent and predictive validity, and was sensitive to environmental and child characteristics. Mothers' reports of language could be useful in large-scale programs.