<![CDATA[This study examined differences in undergraduate EFL students’ essay-writing performance following writing instruction supported by Gemini, Grammarly, and conventional instruction. A quantitative quasi-experimental design was conducted with 91 fourth-semester students in the English Language Education Study Program at Universitas HKBP Nommensen Pematangsiantar during the 2025/2026 academic year. Participants were assigned to three intact groups: Gemini (n = 30), Grammarly (n = 32), and conventional instruction (n = 29). Data were collected through pre-test and post-test essay-writing tasks and assessed using an analytical rubric covering content, organization, vocabulary, language use, and mechanics. Data analysis involved descriptive statistics, assumption testing, a general linear model with the Group × Pre-test interaction, and supplementary post-test comparisons. All groups showed improvement from pre-test to post-test, with the Gemini group achieving the highest post-test mean (M = 81.73) and mean gain (M = 7.60). A one-way ANOVA indicated a significant difference in post-test scores among the groups, F(2, 88) = 6.106, p = .003, ?² = .122. Tukey comparisons showed that the Gemini group significantly outperformed both the Grammarly and conventional groups, while the latter two did not differ significantly. The Group × Pre-test interaction was also significant, F(2, 85) = 4.09, p = .020, partial ?² = .088. These findings indicate that Gemini-supported instruction was associated with stronger essay-writing performance in this context, although the results should be interpreted cautiously given the intact-group design.]]>
Copyrights © 2026
