Developing interdisciplinary methods through AI

Large language models represent individual words, sentences, and larger text elements in the form of a vector in a high dimensional vector space. This is an attempt to characterize the meaning of a text in the form of a single set of numbers and opens for the possibility of comparing meaning in texts and text elements. The use of text embeddings for natural language processing has become a standard tool, but it has yet sparsely been used in education research. In this project, we aim at establishing the use of text embeddings as a reliable and valid research tool for text analysis. We will apply the tool to address (i) the temporal development of the research literature on interdisciplinary education research, with emphasis on the introduction of new methods and approaches over time, (ii) comparisons of courses and curricula across institutions and countries with focus in interdisciplinary education; (iii) studies of interdisciplinary competence and how to measure it from student texts; (iv) analysis of survey and interview using text embedding  methods.