The Effect of Prompt Engineering on Critical Thinking and Writing Skill Quality of High School Students in SMAN 1 Dawarblandong

Abstract

This quasi-experimental research examines the influence of prompt engineering on eleventh-grade students critical thinking abilities and narrative writing quality at SMAN 1 Dawarblandong. Utilizing a pre-test and post-test non-equivalent group design, the study involved 36 students from two intact classes who were assigned to two experimental groups based on their attendance numbers. Group A (n = 18) interacted with ChatGPT using self-constructed prompts, whereas Group B (n+ 18) engaged in narrative writing task using structured prompts developed by the teacher, data were obtained through narrative writing tests assessed with an analytic rubric encompassing five components, content development, organization, vocabulary use, grammatical accuracy, and mechanics. Elements of critical thinking were integrated into the rubric through indicators such as idea originality, logical progression of events, and the relevance of conveyed moral values. The data were analysed using an Independent Sample t-test, which revealed a statistically significant difference in post-test writing performance between the two groups (t = -25.180, 34. P< .001). Students in Group B achieved a higher mean score (M = 82.89) than those in Group A (M = 66.89), indicating a mean difference of 16 points. These results suggest that teacher-designed prompts provide more effective instructional scaffolding than self-generated prompts in AI-assisted writing contexts, underscoring the necessity of explicit instruction in prompt design to enhance EFL student’ writing quality and critical thinking skills. This study offers empirical support for integrating prompt engineering as a structured literacy component in Indonesian secondary education and emphasizes the pivotal role of teacher expertise in facilitating meaningful students’ engagement with generative AI technologies.