國立虎尾科技大學 |

Natural Language Analysis of Student Reflection in Game-Based Learning.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Natural Language Analysis of Student Reflection in Game-Based Learning./
Author:	Carpenter, Dan.
Description:	1 online resource (133 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
Contained By:	Dissertations Abstracts International85-06B.
Subject:	Language. -
Online resource:	click for full text (PQDT)
ISBN:	9798380863902

Natural Language Analysis of Student Reflection in Game-Based Learning.
Carpenter, Dan.

Natural Language Analysis of Student Reflection in Game-Based Learning. - 1 online resource (133 pages)

Source: Dissertations Abstracts International, Volume: 85-06, Section: B.

Thesis (Ph.D.)--North Carolina State University, 2023.

Includes bibliographical references

Reflection is a critical skill for 21st century learners. As learning continues to become increasingly autonomous with the rise of digital learning platforms, such as game-based learning environments, students must effectively regulate their own learning processes to achieve the desired learning outcomes. This self-regulated approach to learning relies on students' ability to engage in meaningful reflection, where they look back on their past learning and usage of learning strategies to evaluate the progress made toward their goals and to inform adaptations for future learning. However, many students do not have strong reflection skills, and these skills are often not explicitly taught in the classroom. Thus, there is a need for educational systems that can help students acquire reflection skills. Integrating reflection into a game-based learning environment presents an opportunity to collect data that captures evidence of students' reflection behaviors during complex learning scenarios that require effective self-regulation to achieve success. Then, leveraging recent advances in machine learning and natural language processing, we can construct automated systems that analyze this data in real time to drive adaptive scaffolds for reflection during game-based learning.This dissertation introduces a natural language processing-based framework for automatically analyzing students' reflections during game-based learning in terms of how effectively they reflect and what they choose to reflect on. The framework leverages recent advances in natural language processing, namely Transformer-based language models, to investigate the reflection behaviors of 153 middle school students as they interacted with CRYSTAL ISLAND, a game-based learning environment for middle school microbiology. First, we explore relationships between students' reflection behaviors and their learning outcomes to empirically evaluate the theoretical benefits of reflection on learning. To this end, we operationalize a measure of reflection depth during science problem solving, extract common topics of students' reflections, align students' reflections with scientific inquiry processes to support a broader exploration of the relationship between students' reflection behaviors and their science problem-solving strategies, and investigate the extent to which students' reflection behaviors change over time as they interact with CRYSTAL ISLAND. Next, we perform contextual analyses that elucidate the relationship between students' learning activities in CRYSTAL ISLAND and their reflection behaviors. Most of the existing work done to analyze students' reflection behaviors deals with reflections written in a post-learning setting, so improving our understanding of the ways in which students' reflection behaviors are related to their learning activities addresses a critical gap in the current literature. Finally, we leverage Transformer-based language models to train robust machine learning models for automatically assessing the depth of students' reflections. Given the small size of our reflection dataset and the high cost of collecting additional reflection data, we explore the use of large language models to synthesize new reflections and find that models trained on synthetic reflections with human-assigned labels significantly improve predictive performance compared to models that only used genuine student reflections.This dissertation presents insights into students' reflection behaviors during game-based learning and the automated reflection analysis framework provides the foundation for AI-enabled adaptive support for reflection during game-based learning.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798380863902Subjects--Topical Terms:

571568
Language.
Subjects--Index Terms:

Digital learning platformsIndex Terms--Genre/Form:

554714
Electronic books.

Natural Language Analysis of Student Reflection in Game-Based Learning.
LDR:05113ntm a22003857 4500 001 1142889
005 20240429063858.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798380863902
035 $a (MiAaPQ)AAI30858001
035 $a (MiAaPQ)NCState_Univ18402041329
035 $a AAI30858001
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Carpenter, Dan. $3 1467410
245 1 0 $a Natural Language Analysis of Student Reflection in Game-Based Learning.
264 0 $c 2023
300 $a 1 online resource (133 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-06, Section: B.
500 $a Advisor: Jiang, Shiyan;Lynch, Collin;Rowe, Jonathan;Lester, James, II.
502 $a Thesis (Ph.D.)--North Carolina State University, 2023.
504 $a Includes bibliographical references
520 $a Reflection is a critical skill for 21st century learners. As learning continues to become increasingly autonomous with the rise of digital learning platforms, such as game-based learning environments, students must effectively regulate their own learning processes to achieve the desired learning outcomes. This self-regulated approach to learning relies on students' ability to engage in meaningful reflection, where they look back on their past learning and usage of learning strategies to evaluate the progress made toward their goals and to inform adaptations for future learning. However, many students do not have strong reflection skills, and these skills are often not explicitly taught in the classroom. Thus, there is a need for educational systems that can help students acquire reflection skills. Integrating reflection into a game-based learning environment presents an opportunity to collect data that captures evidence of students' reflection behaviors during complex learning scenarios that require effective self-regulation to achieve success. Then, leveraging recent advances in machine learning and natural language processing, we can construct automated systems that analyze this data in real time to drive adaptive scaffolds for reflection during game-based learning.This dissertation introduces a natural language processing-based framework for automatically analyzing students' reflections during game-based learning in terms of how effectively they reflect and what they choose to reflect on. The framework leverages recent advances in natural language processing, namely Transformer-based language models, to investigate the reflection behaviors of 153 middle school students as they interacted with CRYSTAL ISLAND, a game-based learning environment for middle school microbiology. First, we explore relationships between students' reflection behaviors and their learning outcomes to empirically evaluate the theoretical benefits of reflection on learning. To this end, we operationalize a measure of reflection depth during science problem solving, extract common topics of students' reflections, align students' reflections with scientific inquiry processes to support a broader exploration of the relationship between students' reflection behaviors and their science problem-solving strategies, and investigate the extent to which students' reflection behaviors change over time as they interact with CRYSTAL ISLAND. Next, we perform contextual analyses that elucidate the relationship between students' learning activities in CRYSTAL ISLAND and their reflection behaviors. Most of the existing work done to analyze students' reflection behaviors deals with reflections written in a post-learning setting, so improving our understanding of the ways in which students' reflection behaviors are related to their learning activities addresses a critical gap in the current literature. Finally, we leverage Transformer-based language models to train robust machine learning models for automatically assessing the depth of students' reflections. Given the small size of our reflection dataset and the high cost of collecting additional reflection data, we explore the use of large language models to synthesize new reflections and find that models trained on synthetic reflections with human-assigned labels significantly improve predictive performance compared to models that only used genuine student reflections.This dissertation presents insights into students' reflection behaviors during game-based learning and the automated reflection analysis framework provides the foundation for AI-enabled adaptive support for reflection during game-based learning.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Language. $3 571568
650 4 $a Computer science. $3 573171
653 $a Digital learning platforms
653 $a Natural language processing
653 $a Game-based learning
655 7 $a Electronic books. $2 local $3 554714
690 $a 0679
690 $a 0984
690 $a 0800
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a North Carolina State University. $3 845424
773 0 $t Dissertations Abstracts International $g 85-06B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30858001 $z click for full text (PQDT)