Validity of caregiver-report measures of language skill for Wolof-learning infants and toddlers living in rural African villages.

Valid indigenous language assessments are needed to further our understanding of how children learn language around the world. We assessed the psychometric properties and performance of two caregiver-report measures of Wolof language skill (language milestones achieved and vocabulary knowledge) for 500 children (ages 0;4 to 2;6) living in rural Senegal. Item response models (IRM) evaluated instrument- and item-level performance and differential function by gender. Both caregiver-report measures had good psychometric properties and displayed expected age and socioeconomic effects. Modest concurrent validity was found by comparing the caregiver-report scores to transcribed child language samples from a naturalistic play session. The caregiver-report method offers a valid alternative to more costly tools, such as direct behavioral assessments or language sampling, for measuring early language development in non-literate, rural African communities. Recommendations are made to further improve the performance of caregiver-report measures of child language skill in these settings.