질적 자료의 정량화 방안

A proposal for analyzing qualitative data quantitatively based on the satisfaction analysis on chatbot-based tasks using sentiment analysis

The present study examines the perception of undergraduate students about chatbot-based English writing tasks, by employing a sentiment analysis approach. To achieve this, the participants were asked to communicate with a chatbot, called Mitsuku, based on seven sub-tasks. The data collected for this study include: 1) English utterances of the participants, generated through the chatbot-based tasks; 2) open-ended responses of the participants in Korean to the chatbot-based tasks; and 3) results from two 5-point Likert-scale questionnaire items, assessing the perceived effectiveness by the participants of chatbot-based tasks for their English learning, as well as their interests in those tasks. The results revealed that the participants were generally satisfied with the chatbot-based tasks. The results from the different tools were moderately or highly correlated overall, and the data of the participants’ English utterances, obtained from the sentiment analysis, was the most significant predictor of their expectations for the effectiveness of chatbot-based tasks for their English learning. The present study provides methodological implications for the employment of the sentiment analysis as an alternative or supplement for the traditional Likert-scale items.