Skip to main content

Artificial Intelligence for use in large scale qualitative data processing

AI head

Project Team members:


Philippa Petts,1 Helen Cramman,2 Rachel Staddon,2, Peter Swift,1 Jacquie Robson,3 Noura Al Moubayed,4 Christian Fischer,5 and Michael Fox6

1Department of Physics, Durham University, Durham, UK

2School of Education, Durham University, Durham, UK

3Department of Chemistry, Durham University, Durham, UK

4Department of Computer Science, Durham University, Durham, UK

5Hector Research Institute of Education Sciences and Psychology, University of Tübingen, Tübingen, Germany

6Department of Physics, Imperial College London, London, UK


The project seeks to investigate the challenges and opportunities for Artificial Intelligence (AI) in facilitating the use of large-scale qualitative data for Higher Education Physics teaching and research. ​​​​​​​




​​​​​​​Big data, commonly characterised through the high quantity of data (volume), speed of data collection (velocity) and variety of modes and timescale of data is becoming increasing available from educational contexts and presents an exciting but complex challenge for education research and practice. Improvements in recording technology and advancement in digital learning environments mean that it is now possible to relatively unobtrusively collect large volumes of high-quality qualitative data quickly from environments where it was previously difficult to capture. Advances in automated speech recognition and artificial intelligence (AI) technologies, such as Deep Learning and Reinforcement Learning have also increased the potential for large-scale automated processing of qualitative data from the complex environments presented by educational settings.


There are, however, important ethical considerations for the automated, or semi-automated processing of large-scale qualitative data, especially if the resulting findings are to be used for decision making e.g. student assessment or policy making. When considering qualitative data, the context in which the data is collected is a vital element within the analysis. The risk when analysing big data is that the context of the data is lost. The potential separation of the direct link between student and teacher in the generation of feedback also breaks a key element of the best practice of producing effective feedback. Subjects with a practical element, such as physics, present a particularly interesting challenge in providing effective feedback within experimental work in the laboratory. The purpose and limit of data processing, the risk of re-identification of participants, bias, profiling and ensuring limitations on data sharing, have all been raised as potential operational risks of AI-based solutions.


Data Collection


The project will collect data from interviews with Higher Education physics teaching professionals from the UK, Europe, Australia and the USA, who have experience in using Artificial Intelligence.  This will then be analysed thematically to assess as to the perceived opportunities and challenges for the use of large-scale qualitative data to support high-quality undergraduate teaching.