Electronic Theses and Dissertations

Date

2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Psychology

Committee Chair

Frank Andrasik

Committee Member

Arthur Graesser

Committee Member

Faruk Ahmed

Committee Member

Xiangen Hu

Abstract

In conversation-based intelligent tutoring systems (ITS), the student is asked challenging questions and there is a multiturn conversation between the computer tutor and the student in constructing an answer to the question. The initial student response after the question provides some information on what the student knows, whereas follow-up contributions of the student presumably extract more of what the student knows. This dissertation investigates the incremental value of follow-up dialogue moves in a conversation-based ITS, specifically ElectronixTutor, a prototype system that has been used in U.S. Navy electronics training. The research addresses two key questions: (1) What is the added value of follow-up dialogue moves in assessing student knowledge beyond initial responses? and (2) How do prior knowledge, student conversational performance, and curriculum-based test outcomes interact to predict the effectiveness of follow-up responses? The dataset that was analyzed consisted of 139 naval trainees on electronics who answered 34 questions with multiturn dialogues. Results show that follow-up dialogue moves significantly improved the assessment of student knowledge. Further, all learners, regardless of prior knowledge, benefited from these follow-up interactions. The study’s implications extend to ITS design, highlighting the importance of incorporating personalized, adaptive feedback based on prior knowledge and leveraging dialogue moves as core features for tracking knowledge growth. These results contribute to the broader understanding of how ITS systems can emulate human tutoring strategies to improve educational outcomes across various domains.

Comments

Data is provided by the student.

Library Comment

PDF

Notes

Open access.

Download

Share

COinS