The TalkMoves Dataset: K-12 Mathematics Lesson Transcripts Annotated for Teacher and Student Discursive Moves

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 4654–4662 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 4654 The TalkMoves Dataset: K-12 Mathematics Lesson Transcripts Annotated for Teacher and Student Discursive Moves Abhijit Suresh1,2, Jennifer Jacobs2, Charis Harty2, Margaret Perkoff1, James H. Martin1,2, Tamara Sumner1,2 1Department of Computer Science, 2Institute of Cognitive Science University of Colorado Boulder Firstname.Lastname@colorado.edu Abstract Transcripts of teaching episodes can be effective tools to understand discourse patterns in classroom instruction. According to most educational experts, sustained classroom discourse is a critical component of equitable, engaging, and rich learning environments for students. This paper describes the TalkMoves dataset, composed of 567 human-annotated K-12 mathematics lesson transcripts (including entire lessons or portions of lessons) derived from video recordings. The set of transcripts primarily includes in-person lessons with whole-class discussions and/or small group work, as well as some online lessons. All of the transcripts are human-transcribed, segmented by the speaker (teacher or student), and annotated at the sentence level for ten discursive moves based on accountable talk theory. In addition, the transcripts include utterance-level information in the form of dialogue act labels based on the Switchboard Dialog Act Corpus. The dataset can be used by educators, policymakers, and researchers to understand the nature of teacher and student discourse in K-12 math classrooms. Portions of this dataset have been used to develop the TalkMoves application, which provides teachers with automated, immediate, and actionable feedback about their mathematics instruction. Keywords: Classroom transcripts, K-12 Education, Accountable Talk, Dialog Acts 1. Introduction Recordings of classroom activities - including video, audio, and transcripts - provide essential data sources for understanding contemporary classroom dynamics and pedagogies (Major and Watson, 2018)(Xu et al., 2018), and for training new forms of intelligent educa- tional technologies. Unfortunately, due to practical and privacy concerns, classroom recordings often have lim- ited availability and sharing is very restricted (Derry et al., 2010). This lack of access to valuable resources and shared data sets has hindered the field’s ability to repli- cate and build on published work. Here, we release and introduce a new corpus: the TalkMoves dataset of over 500 K-12 mathematics lesson transcripts enriched with annotations that capture important, research-based as- pects of classroom discourse. There is widespread agreement that student under- standing is strongly enhanced through sustained inter- action within a learning community and that content- rich discussions should be a prominent and norma- tive feature within K-12 classrooms (Franke et al., 2015). In an effort to delineate specific practices that teachers can use to orchestrate the types of discus- sions encouraged by the Common Core State Stan- dards for Mathematical Practice (Association and oth- ers, 2010), researchers developed an approach to class- room discourse called “accountable talk” (O’Connor et al., 2015). At the heart of accountable talk is the notion that teachers should organize discussions that promote students’ equitable participation in a rigorous learning environment where their thinking is made explicit and publicly available to everyone in the …

Content truncated to 500 words