Summarizing responses to open-ended questions in educational surveys: A microservice-oriented approach

Abstract

Understanding the satisfaction of students’ academic experience is essential for enhancing the educational process and lowering attrition rates. In this context, universities conduct surveys via questionnaires, including both closed-ended and open-ended questions, to track and assess students’ feedback. While closed-ended questions seem useful for rapid quantitative analysis, open-ended ones usually obtain deeper insights and capture a wide spectrum of student experiences. This paper proposes an automated hybrid text summarization approach to handle the problem of assessing vast amounts of qualitative data from responses to open-ended questions. Data preparation, anonymization, normalization, sentence ranking, sentence clustering, extractive, and abstractive summarization, are the main tasks of the proposed approach, which is based on a Microservice-Oriented Architecture. Such an architecture ensures flexibility and scalability, enabling the easy integration of new components that may enhance the approach's overall functionality. Moreover, the applicability of the proposed approach is demonstrated through a dataset containing written responses in the modern Greek language from an online questionnaire completed by about 20.000 students attending study programs at the Hellenic Open University (HOU). The results show that automated text summarization proved to be efficient regarding the time required to produce reports, while also maintaining satisfactory quality.