Applying Natural Language Processing to Explore the Development of Automated Scoring and Feedback Models for Multi-Speaker Science Discussions

Many components of education, including those to support teacher and student learning, are now online and fully digital. But to make productive use of these data sources, the field needs to have a better understanding of what teachers and students do and how they respond to online assessments designed to build and measure complex constructs. This need is especially critical in teacher education where innovative online assessments have become more prominent within teacher licensure and within teacher education settings to support teacher learning. In this project, the intern will identify and investigate meaningful patterns from data generated from teachers facilitating science discussions with student avatars in a simulated classroom to develop and refine automated scoring and feedback models for multi-speaker discussions. The intern will benefit directly by expanding and applying his NLP expertise to examine approaches for supporting teacher learning, as well as collaborating with researchers at ETS.

Faculty Supervisor:

Carl F. Falk

Student:

Partner:

ETS Canada

Discipline:

Sociology

Sector:

Education; Professional, scientific and technical services

University:

McGill University

Program: