Q방법론을 활용한 챗봇 이미지 유형에 대한 사용자 인식 연구

Abstract

Chatbots are expected to have different images depending on how their components are expressed, research on user perception of chatbot images should be conducted first. This study attempted to analyze chatbot image types using Q-Methodology to examine user perception of chatbot images. The research design was conducted in three stages after forming the Q population and selecting the Q sample. The first sample consisted of 100 items, the second sample by factor consisted of 70 items, and the third sample was conducted through questions & answers from experts and in-depth interviewers. Finally, there were 8 items about the chatbot profile image expression, 9 items about the chatbot conversation style, 8 items about the chatbot personality, and 4 items about interaction, in total. A sample of 29 items was extracted. Statements and in-depth interviews were conducted with 40 subjects, men & women in their 20~30s. As a result, chatbot images were classified into four types, focusing on the standard scores of positive and negative statements by type. Type 1 is a friend, Type 2 is a secretary, Type 3 is a leader, and Type 4 is a counsel. This study examined the process of discovering users’ needs, desires, and motivations for chatbot images through Q-Methodology, and served as a tool to classify the perception and type of chatbot images. In the future, we plan to use Q-Methodology to analyze chatbot users’ perceptions in a multidimensional way and use it to establish differentiated strategies for providing chatbot services and interface design direction according to customized chatbot images.