A Study on the Impact of Automated Essay Scoring System Feedback on L2 Writing Quality of Chinese College Learners: A Visual Approach

This study employs text quantitative indicators provided by Natural Language Processing (NLP) technology to conduct a visual analysis of the second language (L2) writing quality of Chinese college learners under the intervention of an Automated Essay Scoring (AES) System from four dimensions—lexical diversity, syntactic complexity, accuracy, and fluency—using eight language feature factors. Visualization techniques such as heatmap, density plot, and box plots were utilized to reveal the interplay between these linguistic features and L2 writing quality. The findings reveal that Average Word Length (AWL), Word Difficulty Average (WDA), and Document Token Count (DTC) are the primary factors influencing AES scoring. The intervention of AES has had a positive effect, as evidenced by improved performance in grammar and spelling accuracy in L2 essays after AES intervention. Meanwhile, challenges remain in the intervention of AES, as L2 writing in terms of lexical diversity, syntactic complexity, and fluency remain in a medium-low range post-intervention. There is considerable variability in vocabulary diversity and syntactic complexity, suggesting a need for more targeted manual instruction and guidance. This study provides a new perspective for exploring the linguistic features and L2 writing quality of Chinese college learners, paving the way for the development of teaching intervention and assessment models.