Discursive resources and agency in classroom discourse: A mixed methods natural language processing approach

This study employs a mixed methods approach to examine science classroom discourse, focusing on the integration of students' discursive resources and the distribution of agency among participants. Utilizing natural language processing (NLP) techniques, we analyzed science classroom videos to identify patterns in the use of everyday versus academic language and to measure the dynamics of teacher-student interactions. Guided by hybridity theory, which emphasizes the blending of students' cultural and academic knowledge, our findings highlight the presence of hybrid discourse spaces that support inclusive and equitable science learning. Qualitative analyses complement the NLP results, providing nuanced insights into how teachers and students navigate and co-construct these hybrid spaces. The study underscores the potential of NLP in educational research to scale the analysis of complex classroom interactions and to inform practices that promote student engagement and equity in science education.