A Method for Extracting Multifaceted Information from Free Descriptions in Questionnaires

Abstract

In Japanese universities, student questionnaires completed in class are used to collect various types of information with two main objectives: students’ impressions of classes and finding areas of improvement. The advantage of a class evaluation questionnaire with numerous questions is that detailed information can be collected; however, contrarily, the number of questionnaires collected and the rate of collection can decrease. This study proposes and tests university students a method to extract relevant information from a simple questionnaire. The proposed questionnaire comprises only five items on students’ impressions of the classes and free descriptions. Text mining techniques were used to analyze the free descriptions and extract the positive or negative feelings expressed by the students. The results were classified according to the questionnaire score. This analysis method efficiently extracts information to identify the factors affecting students’ comprehension, achievement, and satisfaction, which can improve classes. Text mining techniques were used to analyze the free descriptions and extract the positive or negative feelings the students expressed.