Discussion Tracker: Development of Human Language Technologies to Improve the Teaching of Collaborative Argumentation in High School

点击翻译按钮获取中文摘要

Collaborative argumentation - or the building of evidence-based, reasoned knowledge and solutions through dialogue - is essential to individual learning as well as group problem-solving. Student-centered discussions and extended student talk during collaborative argumentation are indicators of robust learning across disciplines. The ability to engage in collaborative problem-solving is also a defining characteristic of 21st century workplaces and civic engagement. However, teaching collaborative argumentation is an advanced skill that many teachers struggle to develop. The goal of this project is to develop an innovative technology called Discussion Tracker, a computer-based system for high school English teachers that uses recent advances in human language technologies (HLT) to provide teachers with automatically generated data and instructional guidance on the quality of students' collaborative argumentation in their classrooms. Discussion Tracker will provide teachers with visual representations of the significant features of their students' collaborative talk and tools for instructional reflection and future planning. The project will improve the teaching and learning of collaborative argumentation in high schools so that students will be prepared for collaborative problem-solving in future educational, workplace, and civic settings.This project leverages recent advances in human language technologies (HLT), data visualization/ analytics, and teacher learning to advance technology that provides automated feedback on classroom talk with the goal of improving teaching effectiveness and student achievement. It will develop novel HLT methods for detecting three significant features of students' collaborative talk: argument moves (claim, evidence, reasoning), specificity, and collaboration (e.g., building on, probing or challenging others' ideas). During years 1 and 2, a series of experiments will be conducted to test Discussion Tracker interface options, to explore teacher learning in response to different types of instructional guidance, and to improve the functionality of the system. Simultaneously, student talk corpora (both existing and collected from the experiments above) will be used to computationally model the three features of student talk by combining relation coding, advances in argument mining, handcrafted features and neural network models with multi-task training. These models will then be incorporated into the Discussion Tracker system to promote teachers' rapid, data-driven reflective practice. In year 3, a large classroom experiment will be conducted to determine the effects of the fully automated Discussion Tracker on teacher learning and instruction across time and school contexts. Generalizable insights will be identified at every stage of the project to promote transferability to future technologies aimed at detecting features of student talk and to teacher learning in similar content areas, grade levels, and language-based learning platforms.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.