Skip to main content Skip to navigation

Identifying Editor Roles in Argumentative Writing from Student Revision Histories

Project Overview

The document explores the application of generative AI in education, particularly in enhancing students' argumentative writing through an intelligent revision assistant. By employing a topic modeling algorithm to analyze students' revision behaviors, the research identifies different editor roles based on three key aspects: operation, purpose, and position. This approach aims to provide personalized feedback that can support students in improving their writing skills. The findings suggest that understanding students' revision processes and the roles they adopt can lead to more effective instructional strategies and tools, ultimately fostering better writing outcomes. The integration of generative AI not only enhances the learning experience but also empowers students to engage more deeply with their writing, promoting critical thinking and self-reflection in the revision process.

Key Applications

Intelligent writing tools that provide localized feedback on text characteristics

Context: High-school and college students engaged in argumentative writing

Implementation: Using topic modeling (LDA) on revision histories to identify editor roles based on operation, purpose, and position of revisions

Outcomes: Identified editor roles correlate with writing improvement, especially the Persuasive editor role.

Challenges: Limited understanding of how each role contributes to writing success; requires validation of revisions and their purposes.

Implementation Barriers

Technical barrier

Complexity of accurately modeling and validating the editor roles in student writing revisions.

Proposed Solutions: Future plans include using a Markov model to consider revision order and expanding the vocabulary for more comprehensive feedback.

Project Team

Tazin Afrin

Researcher

Diane Litman

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Tazin Afrin, Diane Litman

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies