Objectives This study aims to explore the potential of using ChatGPT as an alternative to address the difficulties faced by freshmen in university writing classes and empirically verify its impact on writing efficacy and academic achievement. Methods The study involved 130 students from a four-year university in Gyeonggi Province enrolled in a Basic Writing course. After the midterm exam, students were divided into an experimental group using ChatGPT and a control group that did not. A pre- and post-survey on writing efficacy was conducted, and data on academic ach ievement were collected through midterm and final exam scores, along with qualitative data from students' re flective journals and instructor observation logs. Results The statical and qualitative analysis revealed that students who used ChatGPT showed significant im provement in overall writing efficacy, particularly in expressive functions. By referencing ideas and vocabulary pro vided by ChatGPT, students were able to ease difficulties in generating and expressing their own ideas, expanding their thinking to express themselves more creatively and effectively. In terms of academic achievement, the ChatGPT group performed significantly better on the final exam compared to the control group. Correlation analy sis between writing efficacy and academic achievement showed that grammar and expression functions had a stronger correlation with academic performance than structural functions. Conclusions These results suggest that the samples and personalized feedback provided by ChatGPT improved students' self-editing abilities in writing, boosted their motivation and engagement, and ultimately led to better academic outcomes. This study offers empirical evidence that AI tools like ChatGPT can enhance creative ex pression, writing efficacy, and academic performance by providing individualized feedback in writing education. It also offers valuable insights into effectively integrating AI tools in educational settings.
Read full abstract