Quantitative interviews: extending transcript analysis using natural language processing

When the COVID-19 pandemic hit the world, it affected education at all levels. The shift to wholly online delivery has been challenging for higher education students and staff. However, it has also led to opportunities for new methods to deliver learning and teaching online. During this period, mathematics and statistics support also shifted their services online. In this study, students and tutors were interviewed to understand the opportunities and challenges they encountered with wholly onlinelearning, teaching and support. Twenty-three participants were selected from University College Dublin, Ireland, and Western Sydney University, Australia and one-on-one interviews were conducted in late 2020. While interviews are an excellent way to gather detailed information, analyzing them usually requires qualitative techniques which can be time-consuming and result in a small sample size. In this study, we aim to identify common themes around online mathematics and statistics support by Natural Language Processing (NLP). Interview transcripts were converted to numerical data using text mining techniques and classification and topic modelling methods were applied to identify common themes in the transcripts via the R programming language. These findings were compared to the previous qualitative study results to investigate how software-based models perform versus human-basedmodels. Implementing NLP techniques can help to increase sample size, reduce project time and costs.