ANA JEAN E. CARDONA , MARIA LUZ T. VILLANUEVA , ELLANIE C. TEMPLE
DOI: https://doi.org/This study examined the impact of AI-assisted prewriting on the writing competence and perceptions of higher education students. A total of 90 participants engaged in structured prewriting activities supported by ChatGPT, designed to enhance idea generation, paragraph organization, and sentence construction. Writing competence was assessed across three pre- and post-intervention tests, evaluating Content Development, Coherence, Grammar, Creativity, and Overall Writing Scores. Paired-sample t-tests demonstrated statistically significant improvements across all sub-dimensions (e.g., Overall Score: Pretest M = 65.65, SD = 8.64; Posttest M = 73.67, SD = 7.67; t = 11.10, p < .001). Normality tests confirmed the suitability of parametric analyses. Students’ perceptions of ChatGPT were measured across three phases using a structured questionnaire. Repeated measures ANOVA indicated significant gains in Perceived Usefulness (F = 12.45, p < .001), Perceived Ease of Use (F = 7.32, p = .001), Ethical Awareness (F = 5.68, p = .004), and Overall Perception (F = 10.21, p < .001), with moderate effect sizes. Qualitative analysis of reflective journals and focus group discussions (n = 15) revealed four key themes: Cognitive Support in Prewriting, Motivation and Confidence, Ethical Awareness and Responsible Use, and Digital Literacy Challenges. The integration of quantitative and qualitative findings highlights ChatGPT’s dual role as a cognitive and motivational scaffold in writing instruction. Results suggest that AI-assisted prewriting can enhance writing performance, foster responsible AI use, and develop digital literacy, underscoring the need for structured guidance and ethical integration in higher education pedagogy.
