Harnessing Micro-Topics Arranged in Learning Pathways for Spaced Retrieval, Reading, and Collaborative Note-taking

Abstract

This dissertation explores the utilization of a Knowledge Graph of ``Micro-topics'' arranged into ``Learning Pathways'' (KGMLP) through three learning activities: spaced retrieval, reading, and collaborative note-taking. The objective is to enhance four (out of six) levels of Bloom's digital taxonomy of cognitive processes: remembering, understanding, analyzing, and evaluating.

The research initiates with the introduction of a novel spaced retrieval practice tool, crafted to synchronize with a course's existing knowledge graph. Chapter 2 unveils a two-fold study; firstly, it examines the correlation between students' engagement with this tool and their academic performance, revealing significant improvements in exam scores. Secondly, it scrutinizes the ``counting days'' incentive scheme, designed to encourage consistent tool usage. Through a semester-long randomized controlled trial (RCT) within a class and a separate RCT involving 71 instructors, findings demonstrate the efficacy of this mechanism in enhancing students' motivation for spaced retrieval, which subsequently leads to improved academic outcomes, particularly notable in students with lower GPAs.

Chapter 3 transitions to the structural representation of KGMLP, debating whether to emphasize the visualization of relationships between entire paragraphs or to dissect paragraphs into shorter phrases for detailed relational visualization. A meticulously conducted experiment with 419 participants showcases the superior efficacy of KGMLP formats that encapsulate entire paragraphs, outperforming traditional Novakian concept maps in enhancing both factual and inferential comprehension and recall capabilities, immediately, as well as in subsequent evaluations conducted three days and a week later. Feedback analysis indicates a strong preference for visual connections between complete paragraphs, with an emphasis on having all content accessible on a single page.

Chapter 4 explores the creation and refinement of KGMLPs within a collaborative framework. The introduction of ``1Cademy,'' a pioneering platform, marks a significant stride in enabling social learning through collaborative note-taking over extended periods. This chapter addresses the challenges inherent in such collaborative environments, including motivating student engagement in peer content refinement and consensus building in micro-topics and learning pathways. Four distinct design elements are put forward and empirically tested: a proposal system for content modification with a voting-based approval process, a mandatory inclusion of a learning pathway in new node proposals, and a novel community-specific contribution scoring system. Through an extensive longitudinal study and field trials, the research demonstrates the adaptability of students to these design features, ensuring effective collaboration without the fragmentation or privatization of the knowledge graph. 1Cademy has successfully engaged 1,754 students and researchers from 208 institutions since its inception four years ago, resulting in a comprehensive KGMLP consisting of 50,947 nodes and 271,836 proposals.

This dissertation presents a multifaceted exploration of KGMLP in educational contexts, highlighting its potential to enhance key cognitive processes and foster a collaborative learning environment. The studies and results within offer significant insights into the effective integration and application of KGMLP in modern educational settings.