THE EFFECT OF USING GRAMMARLY AS AN AI GRAMMAR CHECKER ON THE GENERAL WRITING OF THE ENGLISH EDUCATION DEPARTMENT STUDENTS

Abstract

Writing is a demanding productive skill for EFL learners, and grammatical accuracy is among their most persistent difficulties. AI-based grammar checkers such as Grammarly are widely used, yet empirical evidence of their effect in local university contexts remains limited. This study examined the effect of Grammarly, an AI grammar checker, on the recount text writing of English Education Department students at UIN Alauddin Makassar. A pre-experimental one-group pre-test and post-test design was employed. Twenty-eight third-semester students of one class (PBI C) were selected through purposive sampling. Data were collected through a recount writing test scored by two raters with Jacobs et al.'s analytic rubric and complemented by semi-structured interviews with three students. Descriptive statistics and the Shapiro–Wilk test were followed by the Wilcoxon Signed-Rank Test in SPSS 26 because the post-test scores were not normally distributed. The mean score rose from 65.00 (SD = 13.033) to 73.57 (SD = 9.319), and the difference was significant (Z = −3.353, p = .001). All five components improved, most in language use (+2.61) and organization (+2.43) and least in mechanics (+0.50). The interviewees viewed Grammarly positively, although two of them accepted all suggestions without checking them. The findings can be interpreted through Schmidt's Noticing Hypothesis and Mayer's Cognitive Theory of Multimedia Learning. Because the study had no control group and used a test that was not formally validated, the results should be interpreted with caution. Grammarly appears most useful as a supporting tool under lecturer guidance.