Teachers’ Language in Mathematics: A Question of Student Diversity

• In Event: Straight Talk: Natural Language Processing Approaches to Characterizing Student Opportunities to Learn in Mathematics

Abstract

Language is a key factor in classroom settings (Chin, 2006). In particular, what teachers say and how they say it is an important source of information for students. Thus, classroom talk can be essential for students’ learning (Sedova et al., 2018). In mathematics education, language plays a dual role, supporting both conceptual understanding and the acquisition of technical vocabulary (Pöhler & Prediger, 2015). At the same time, research shows that teachers’ instructional moves and language may differ depending on student characteristics such as ethnicity or socioeconomic status (Harris & Rosenthal, 1985). Therefore, we examine the language of teachers with regard to various dimensions, such as mathematical language, or sentiment (e.g., Harber, 2023).

Educational research has rarely systematically examined language biases on the basis of authentic teaching transcripts. Using data from the Global Teaching InSights study (OECD, 2020), we will explore whether teachers' language varies with student diversity, particularly in terms of immigrant background, language spoken at home, and socioeconomic status. The sample includes 50 teachers from 38 German schools and a total of 1140 students. First, we investigate the ratio of teacher talk to student talk, language sentiment, the use of technical language and lexical diversity using Natural Language Processing techniques. These language characteristics are then related to student diversity using hierarchical linear modeling. Preliminary results show that lexical diversity in teachers' language decreases as the number of students with an immigrant background in a class increases. We see similar patterns with mathematical vocabulary and sentiment. Teachers use less mathematical vocabulary and more negative language in classrooms with a greater proportion of students with a migrant background. We observe the opposite pattern for student socioeconomic status: the higher the socioeconomic status of the students in a class, the more math vocabulary is used and the more positive the language is. We find no correlation between the teacher-student word ratio and class composition.

In a second step, particularly contrasting individual cases will be examined in more detail using qualitative video analysis (Martens & Asbrand, 2022). Based on the results of the quantitative language analysis, particularly striking sequences in the classroom will be analyzed qualitatively. Using an interaction-analytical approach, the qualitative analysis will complement the quantitative findings, creating a reciprocal relationship where particularly striking results are contextualized through qualitative insights. Our research will provide unique insights into how teacher speech in mathematics varies with student diversity, thus contributing to national and international research on teacher-student interaction in the context of diversity.

Authors

• Hannah Kleen, DIPF | Leibniz Institute for Research and Information in Education

• Katharina Wehking, University of Osnabrück

• Patrick Schreyer, University of Kassel

• Helene Zeeb, University of Erfurt