NITELITE: Studying Student Language and Teacher Feedback Through Explanation and Asynchronous Collaboration

Erickson, John A.

Etd

NITELITE: Studying Student Language and Teacher Feedback Through Explanation and Asynchronous Collaboration

Public

Online education technologies have provided support for teachers and students in a plethora of ways. Teachers have, and continue, to utilize the automated support; mainly, their automated scoring, feedback messages and student reports. Students benefit from the support of automated scoring, immediate feedback, common wrong answer messages, hints and scaffolding. However, most of the support has been limited to questions with structured answers. Mainly, the support for these sys- tems are widely limited to multiple choice or fill in the blank questions. Questions which have well defined answers. While these provide insights into the students learning and performance; understanding what the mistake was, or why, is based on cumulative common wrong choices. Not every student will process information the same way. While there are many common mistakes, why the student is making this mistake is unique to them. For teachers, the language a student uses provides deep insight into the students’ process of thinking, where they may be struggling in the material, or what steps specifically may be causing the confusion. Requir- ing students to elaborate their work, and write through their steps taken, assists the teacher in identifying each student’s unique understanding of the materials and requires a larger range of cognition from the student. For the student, the direct communication through teacher feedback in open response questions provides them with a more personalized explanation to why they scored what they did. However, for a teacher to score open responses and reply to students, it’s inevitably a tedious task. My research aims to utilize student language and teacher feedback to: 1. Develop open response support for teachers with a set of tools which utilize natural language processing to automatically grade and suggest feedback for student answers to open response questions. 2. Evaluate the potential unfairness of these predictive models. 3. Diversify this open response feedback by evaluating the sentiment teachers use in their feedback. 4. Develop open response support for students and evaluate its effectiveness with a randomized controlled trial on NITELITE (Nonsynchronous Integrated Technology Environment - Learning from Interdependent Terminologies and Explanations), a tool for utilizing open response rationale for asynchronous collaborations within intelligent tutoring systems.

Creator