Survey of 78 economics-education students at Universitas Jambi finds ChatGPT ease of use scores highest (M = 3.12) while accuracy and critical thinking rate lower

The study reports a ranked set of five dimension means for students' perceptions of ChatGPT in Educational Economics learning at Universitas Jambi: ease of use highest (M = 3.12), then knowledge (M = 3.02), satisfaction (M = 2.94), motivation (M = 2.83), and activity (M = 2.81). Prior work on perceptions of generative AI in education often addresses general higher-education settings or language courses; this study anchors the measurement in a specific course, Educational Economics, and a specific institutional setting, providing dimension-level descriptive data for that context. A cross-sectional questionnaire with total sampling of 78 students, 16 items on a four-point Likert scale, analyzed with descriptive statistics using mean scores and predetermined categories; the text reports no reliability coefficients, significance tests, or effect sizes.

Students perceived ChatGPT as accessible and helpful for understanding course materials and obtaining rapid responses, yet perceptions were comparatively lower regarding response accuracy, critical thinking, and motivation for academic writing. The result places positive overall perception alongside reservation about higher-order learning activities, indicating that endorsement of the tool is not the same as endorsement of its role in higher-order cognitive tasks. Derived from comparisons of dimension means within the same self-report questionnaire, reflecting subjective perception rather than objective learning outcomes; the text provides no item-level distributions or statistical tests.

The authors conclude that appropriate pedagogical guidance is needed to encourage critical evaluation of AI-generated information, independent learning, and responsible use of ChatGPT in Educational Economics learning. It converts descriptive perception data into a direction for teaching practice, linking tool use to critical evaluation and responsible-use concerns. An author recommendation grounded in descriptive results; the text reports no intervention implemented or effect verified.

The study applies to the specific setting of students taking the Educational Economics course in the Economic Education Study Program at Universitas Jambi in the 2025/2026 academic year. Its value is as a course-level baseline description from the student perspective: instructors can use it to judge where additional guidance may be needed, such as accuracy checking, critical-thinking training, and academic writing. Researchers and teaching administrators interested in how students in a particular institution and course self-assess ChatGPT's role can draw on this descriptive data directly; readers seeking cross-institution or cross-discipline comparisons, or causal conclusions, would need to combine it with other studies.

The loaded text is an incomplete version containing only the abstract and reference list, without the body's method details, the content of the dimension items, scale reliability and validity information, data collection timing, or ethics procedures, and without figures or statistical tests. It is therefore not possible to judge whether differences among the five dimension means are statistically meaningful, nor to verify the specific thresholds of the 'predetermined categories.' In addition, the perception data are self-reported, so the relationship to actual learning outcomes remains an open question, and the single-institution, single-course sample means the scope of the conclusions awaits testing in further settings.