An Accompaniment-Oriented Human–Machine Collaborative Model for EFL Writing Assessment

This study constructs and implements an accompaniment-oriented human–machine collaborative model for EFL writing assessment, which integrates both process-oriented and product-oriented assessment through the use of technological tools. The model further clarifies the respective roles of teacher assessment, self-assessment, peer assessment, and machine assessment, thereby establishing a comprehensive and systematic framework for EFL writing evaluation. By incorporating the evaluation of EFL writing learning strategies, EFL writing strategies, and EFL writing outcomes, the model extends beyond existing studies in terms of scope, diversity of assessors, and richness of methods. Moreover, it combines multiple complementary approaches—including questionnaires, think-aloud protocols, keystroke logging, screen-recording software, corpus-based data analysis, and multi-faceted Rasch modeling—to provide triangulated evidence, thereby enhancing the overall reliability of the proposed assessment model.