<![CDATA[This study investigated the impact of ChatGPT as a guided thinking partner on ninth-grade students’ recount text writing skill at SMP Negeri 2 Suwawa. Although previous studies have shown that ChatGPT can support writing instruction, limited classroom-based quantitative evidence has examined its supervised use as a thinking partner rather than as an automatic text generator in junior secondary EFL recount writing. Grounded in sociocultural scaffolding theory, this study positioned ChatGPT as guided cognitive and linguistic support during idea generation, outlining, drafting, revising, and editing. A quantitative pre-experimental one-group pre-test and post-test design was employed with 30 students selected through purposive sampling. Students’ recount texts were assessed using an analytic rubric covering structure, grammar accuracy, vocabulary use, and coherence. The data were analyzed using descriptive statistics, the Shapiro-Wilk test, the Wilcoxon Signed-Rank Test, and effect size calculation in IBM SPSS Statistics 27. The findings showed that the mean score increased from 72.33 in the pre-test to 85.83 in the post-test, with improvements across all writing indicators, especially coherence and text structure. The Wilcoxon test indicated a significant difference between pre-test and post-test scores (Z = -4.102, p < .001), with a large effect size (r = .75). The study concludes that ChatGPT can support recount text writing when it is used ethically, reflectively, and under teacher supervision as pedagogical scaffolding rather than as a substitute for students’ own writing process.]]>
Copyrights © 2026
