Journal Paper Published

Abstract

Purpose:
Although existing studies have begun to probe students’ engagement with Generative artificial intelligence (GenAI) feedback on L2 writing, few capture the full continuum of engagement that arises when students both interact with GenAI and subsequently revise their texts. This study addresses this methodological gap
by developing and validating a comprehensive instrument to assess student engagement in GenAI-mediated writing environments.

Methodology:
A 23-item GenAI feedback engagement scale was constructed to jointly assess behavioral, cognitive, and emotional engagement during the interaction and revision phases. Data were collected from 890 undergraduates across nine universities in China. Exploratory and confirmatory factor analyses were
employed to validate a proposed six-factor, second-order structure, and measurement invariance was examined across demographic variables.
Findings: The results supported the hypothesized six-factor, second-order structure and established measurement invariance across gender, university prestige, and academic major. Cronbach’s alpha coefficients ranged from 0.83–0.91, indicating
strong internal consistency.

Originality/value:
This study extends the theoretical framework of student engagement with feedback to the emerging context of GenAI-supported L2 writing by capturing engagement across both the interaction and revision phases. The validated scale
offers researchers and practitioners a psychometrically robust diagnostic tool for assessing and fostering student engagement in GenAI-mediated writing instruction.

Keywords