Generative AI Prompt Engineering for L2 Writers: The Effects of Teaching Prompting Strategies on Learners’ Writing Performance and Perceptions
CALICO Journal · 01.11.2026 · 43 (3) · 422-452
DOI: 10.3138/calico-2025-0063 ↗Kurz gefasst
Fachartikel in der Zeitschrift CALICO Journal, erschienen am 01.11.2026. Verfasst von Che-Han Chen und 2 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben.
Abstract (Originalsprache)
In second language (L2) writing contexts, research in generative AI (GenAI) has shown that carefully engineered prompts are critical to generating appropriate feedback catered to the needs of individual learners. However, empirical evidence regarding the best practices for integrating the principles and best practices of prompt engineering in real-world L2 writing classrooms remains scarce in the current literature. To address this gap, this classroom-based study explores how ChatGPT prompt engineering instruction affects adult intermediate-level L2 English learners’ writing performance in a semester-long English for Academic Purposes course using an experimental between-groups research design. A method of prompt engineering referred to as the chain-of-thought approach (CoT) was employed for this study. This prompting strategy includes context-setting, tailored feedback requests, and active engagement with AI-generated feedback. During the instruction period, the experimental group received explicit ChatGPT prompting instruction, while the comparison group did not. Data collected includes 4 timed in-class argumentative essays, writing test scores, a perception survey, and semi-structured interviews. Wilcoxon rank-sum test was used to measure students' writing performance between groups. Results showed significant post-test gains for the experimental group in organization, content development, language use, and overall writing quality, compared to the comparison group. Results suggest that while the CoT prompt technique seemed to help them improve their writing proficiency, most participants expressed that teacher intervention remains essential for their L2 writing performance. This research provides insight into the efficacy of GenAI in facilitating the acquisition of L2 writing and AI literacy.
Zitieren
APA
Chen, C., Zheng, Y., & Dixon, D. H. (2026). Generative AI Prompt Engineering for L2 Writers: The Effects of Teaching Prompting Strategies on Learners’ Writing Performance and Perceptions. CALICO Journal, 43(3), 422-452. https://doi.org/10.3138/calico-2025-0063
Vancouver
Chen C, Zheng Y, Dixon DH. Generative AI Prompt Engineering for L2 Writers: The Effects of Teaching Prompting Strategies on Learners’ Writing Performance and Perceptions. CALICO Journal. 2026;43(3):422-452. doi:10.3138/calico-2025-0063
BibTeX
@article{Chen2026,
author = {Chen, Che-Han and Zheng, Yiwen and Dixon, Daniel H.},
title = {Generative AI Prompt Engineering for L2 Writers: The Effects of Teaching Prompting Strategies on Learners’ Writing Performance and Perceptions},
journal = {CALICO Journal},
year = {2026},
volume = {43},
number = {3},
pages = {422--452},
doi = {10.3138/calico-2025-0063}
}
EUVIMED