Conference Presentation: Trait-Based Automated Essay Scoring in Rater Training for the Evaluation of Young English Language Learner’s Writing
Date: Saturday, September 26, 2026 | 4:15 PM CT
Session Format: Poster
Presenters:
Aubrey Sahouria (Center for Applied Linguistics)
Frank Wucinski (Center for Applied Linguistics)
Joshua Smith (Center for Applied Linguistics)
Yu Lan Su (Center for Applied Linguistics)
Hyeonseong Li (Center for Applied Linguistics, University of Maryland)
Description:
Trait-based automated essay scoring (AES) has been integrated into several high-stakes language proficiency assessments to expedite scoring and verify human ratings for writing. Both AES engines and human raters require extensive training to reliably assign scores across multiple traits, especially when assessing younger English language learners’ (ELLs; grades 3-5) writing skills, which demand developmentally sensitive evaluation. Assessing language proficiency in young children presents unique challenges due to developmental variation in attention, cognition, emotional regulation, and verbal expression. Young children may have limited attention spans, fatigue quickly, or struggle with unfamiliar testing formats. Training raters for this context requires careful review and selection of benchmark writing samples across a range of criteria, which can be resource-intensive, especially when accounting for young ELL’s developing skills. Recent advances in population-specific, trait-based AES methods present an opportunity to streamline the selection of representative writing samples for rater training, potentially reducing resource-intensity and diversifying training options. This literature review explores and summarizes the utility of tailored, trait-based AES for the selection of rater training writing samples in high-stakes English language proficiency assessments for younger learners.
