MEASURING AND UNDERSTANDING STUDENTS’ SELF-REGULATED LEARNING IN TEXTUAL DATA IN COMPUTER-BASED LEARNING ENVIRONMENTS

Self-regulated learning (SRL) describes the process in which learners take active control of their learning by regulating their attention and effort in pursuit of goals. Students who are skilled in SRL are able to effectively set goals, search for information, and direct their attention and cognitive resources to align their efforts with their objectives. These skills are especially important in computer-based learning environments (CBLEs), where students learn in an individualized manner and are expected to take active control deciding what to learn, when to learn, and how to learn. Prior work has used log data to measure SRL in real time, identifying when students use or fail to use SRL and informing adaptive scaffolds that support self-regulation. However, most research has relied on behavioral logs (such as clicks and actions) and only a few studies have considered the textual data for this purpose, which include data that records students’ written responses or thinking process in a textual format. In this dissertation, I examine the use of textual data to measure and understand SRL, contextualized within computer-based learning environments for individual learning. Specifically, the first study explores the potential for identifying SRL indicators in open-ended responses in CueThink, a digital learning platform for middle school mathematics. These indicators reflect students’ engagement in cognitive operations such as assembling and monitoring. Follow-up analyses examine the relationships between these SRL indicators and students’ affect, engagement, and performance, providing evidence of the critical role SRL plays in mathematical problem-solving. The second study investigates the measurement of SRL processes in think-aloud protocols during problem-solving in intelligent tutoring systems. SRL processes, including processing information, planning, enacting, and recognizing errors, were detected and analyzed in relation to moment-by-moment performance and system design features. Finally, the third study examines the potential for detecting gaming the system (an ineffective use of SRL) in students’ open-ended responses to self-explanation questions. Through this line of research, this dissertation aims to provide a summary on the possibility of measuring and understanding SRL using textual data, an important yet understudied topic in the field of learning analytics and computer-based learning.

Author dc:creator

• Zhang, Jiayi

Advisor dc:contributor.advisor

• Chen, Bodong

• Education

Language dc:language.iso

en

Repository record dc:identifier.uri

https://repository.upenn.edu/handle/20.500.14332/62353

OAI identifier oai:identifier

oai:repository.upenn.edu:20.500.14332/62353