<![CDATA[The rapid development of Artificial Intelligence (AI) has created new opportunities for supporting programming learning, particularly through the use of ChatGPT and prompt engineering techniques. This study aims to examine the impact of ChatGPT-assisted learning on students’ programming ability and to compare the effectiveness of two prompting techniques, namely Text-to-Code and Code-to-Code. A quantitative approach with a pretest-posttest two-group experimental design was employed. The participants consisted of 35 Computer Education students, divided into a Text-to-Code group (n = 18) and a Code-to-Code group (n = 17). Data were collected through programming skill tests and a student perception questionnaire. The results showed that both prompting techniques improved students’ programming ability at a medium level. The Text-to-Code group obtained a mean N-Gain score of 0.424, while the Code-to-Code group achieved a higher mean N-Gain score of 0.507. The statistical test indicated a significant difference between the two groups (p = 0.035), suggesting that the Code-to-Code technique was more effective in enhancing programming ability. In addition, students in both groups showed positive perceptions toward the use of ChatGPT, with aggregate mean scores of 4.00 for Text-to-Code and 3.94 for Code-to-Code. These findings indicate that ChatGPT can support programming learning effectively, particularly when students are encouraged to write and analyze code before using AI assistance.]]>
Copyrights © 2026
