Developing an Open Multi-Domain Assessment Battery Using Laboratory Tasks for Children in Low- and Middle-Income...

Recent years have seen growing interest in the use of randomized designs to evaluate the impacts of interventions on early child development in low- and middle-income countries. Measuring these impacts requires assessments that can track development prior to school entry in settings that are often noisy and outdoors. Drawing from research in developmental cognitive science, we created an open multi-domain assessment battery for testing children in low- and middle-income countries consisting of tasks based on experimental paradigms used in laboratory settings. These tasks were designed to target abilities that emerge across cultures early in life and predict later performance on related skills in adulthood. This battery was administered to 1435 Ghanaian children whose parent previously took part in a randomized controlled field experiment on the impacts of secondary education on their children's development, measuring performance on tasks at ages 1.5, 3.5, 5.5, and 7.5 years. Starting at age 3.5, children's responses to the tasks were internally consistent and above chance, with overall test results predictive of later performance. There was strong evidence of validity both for tasks in numeracy and spatial reasoning, which were designed to tap into culturally invariant cognitive capacities, as well as those in language and executive functioning, which required more context-specific adaptation. In contrast, there was little evidence of validity in infant tasks or tasks focused on intuitive physics or mental state reasoning. Nevertheless, our results suggest that our approach to assessment design is effective at producing tasks that are internally consistent and sensitive to individual differences within cohort. By making our testing materials and data freely accessible, we hope to provide a foundation for the development of tools for studying cognitive development in field settings.