The purpose of this study is to reveal the characteristics of how college student writers use ChatGPT in the writing process by analyzing the ‘question’ aspect of the writer group according to the writing level. To this end, 331 questions appearing in the <Log using ChatGPT> written by 60 students were analyzed according to the criteria of three question types, as well as the frequency and continuity of questions for each writing process.As a result, ⑴ college writers showed a bias in their question patterns concerning the process of ‘creating content’, asking ‘one-time questions’ rather than ‘continuous questions’, ‘dependent questions’ rather than ‘initiative questions’, and ‘general questions’ rather than ‘specific questions’. In terms of functions required of ChatGPT, many questions arose in the order of ‘information (presentation)’, ‘evaluation’, ‘creation’, ‘application’, and ‘analysis’. ⑵ The question patterns of the upper and lower groups according to the writing level contrasted greatly. The upper group used ‘continuous’, ‘initiative’, and ‘specific’ questions relatively more, and showed a tendency to ask various questions in terms of request functions for ChatGPT. On the other hand, in the subgroup, attention was paid to the problem of the unethical use of ChatGPT due to the ‘invention’ question along with the contrasting aspect.Based on the above data, this studey indicated four areas that future ‘Question Education’ should take into consideration. First, in the field of writing education, the direction of ‘question education’ should be set under the goal of ‘priority improvement of learners' writing ability’. Second, prior to education on question methods, various types of questions should be taught. Third, useful question types and question sequence strategies for each writing process should be developed in connection with existing writing education contents. Finally, an agreement between the teacher and the learners concerning the use of ‘invention’ questions must be drawn up in a very specific manner.
Read full abstract