Assessing Development across Cultures: Invariance of the Bayley-III Scales across Seven International MAL-ED Sites

The Bayley's Scales of Infant and Toddler Development-Third Edition (Bayley-III) were used to measure the development of 24-month-old children (N = 1,452) in the Interactions of Malnutrition and Enteric Infections: Consequences for Child Health and Development (MAL-ED) study (an international, multisite study on many aspects of child development). This study examined the factor structure and measurement equivalence/invariance of Bayley-III scores across 7 international research sites located in Bangladesh, Brazil, India, Nepal, Pakistan, Peru, and South Africa. Exploratory and confirmatory factor analyses were used to identify the factor structure of Bayley-III scores. Subsequently, reliability analyses and item response theory analyses were applied, and invariance was examined using multiple-indicator, multiple-cause modeling. The findings supported the validity, but not invariance, of Bayley-III language scores at all seven sites and of the cognitive and motor scores at six sites. These findings provide support for the use of scores for research purposes, but mean comparison between sites is not recommended. Impact and Implications: In measurement, validity refers to the extent to which we are measuring what we intend to measure and the appropriateness of inferences we make based on our measurements. The validity of scores from measures often varies across cultures, and this study examined the validity of a measure of child development among children from sites in seven low- and middle-income countries. The findings indicate that the majority of the scores are valid for research, but measurement differences are evident such that it is not appropriate to compare mean scores across sites. [Co-written with the MAL-ED Network Investigators.]