Reducing Test Bias Through Dynamic Assessment of Children's Word Learning Ability

Abstract

This study examined the performance of preschool children from culturally and linguistically diverse backgrounds, both typically developing and with low language ability, on a word-learning task. A pretest-teach-posttest method was used to compare a mediation group to a no-mediation group. Children in the mediation group were taught naming strategies using mediated learning experience (MLE). Results indicated that typically developing and low language ability children were differentiated on the basis of pretest-posttest change and that dynamic measures (e.g., posttest scores of single-word labeling and modifiability ratings from the mediation sessions) predicted the ability groups better than static measures (e.g., pretest scores of single-word labeling, description, and academic concepts). These results suggest that dynamic assessment approaches may effectively differentiate language difference from language disorder.