Metacognitive Strategies, AI-based Writing Self-efficacy and Writing Anxiety in AI-assisted Writing Contexts: A Structural Equation Modeling Analysis

Abstract

Although the impact of Artificial Intelligence (AI) on English writing is significant, EFL learners’ writing anxiety can adversely affect their performance. Much research has explored the individual effects of metacognitive strategies or writing self-efficacy on writing performance. However, there is limited investigation into how these constructs influence writing anxiety, particularly within AIassisted writing contexts. This study aimed to investigate the effects of metacognitive strategies and AI-based writing self-efficacy on writing anxiety among EFL learners, as well as to explore the mediating role of AI-based writing self-efficacy in the interrelationship between metacognitive strategies and writing anxiety by structural equitation modeling (SEM). A total of 193 participants completed questionnaires on metacognitive strategies, AI-based writing self-efficacy, and writing anxiety. The results indicated that metacognitive strategies encompassing planning, monitoring, and evaluating, and AI-based writing self-efficacy have a negative impact on writing anxiety. Specifically, AI-based writing self-efficacy fully mediates the relationship between planning strategies and writing anxiety, while partially mediating the relationships between monitoring and evaluating strategies and writing anxiety. These findings underscored the crucial role of AI-based writing self-efficacy in alleviating writing anxiety among EFL learners in AI-assisted contexts and suggested that enhancing metacognitive strategies can reduce learners’ writing anxiety by boosting their confidence in using AI-based tools.

Keywords

Full text

How to cite