HCII PhD Thesis Proposal: Conrad Borchers
When
-
Where
NSH 3305
Description
Personalized Goal Setting Support to Enhance Technology-Supported Learning
Conrad Borchers
HCII PhD Thesis Proposal
Date & Time: Friday, September 12th @ 3:00 pm EDT
Location: Newell-Simon Hall (NSH) 3305
Zoom Link: https://cmu.zoom.us/j/91838644909 // Passcode: 713904
Committee:
Vincent Aleven (Co-Chair), Carnegie Mellon University
Kenneth R. Koedinger (Co-Chair), Carnegie Mellon University
Geoff Kaufman, Carnegie Mellon University
Roger Azevedo, University of Central Florida
Emma Brunskill, Stanford University
Abstract:
Sustained student engagement remains a persistent challenge in realizing the full potential of AI-enhanced learning environments such as intelligent tutoring systems. Goal setting—defining clear, quantifiable objectives for effort and performance—has been shown to enhance motivation and performance. However, traditional goal-setting interventions in K-12 education, often reliant on teacher, parent, or tutor facilitation through paper-based contracts, face scalability and integration challenges in digital learning contexts.
This thesis addresses this gap by designing and studying technology-mediated, personalized, and data-driven approaches to facilitate and scale goal support, defined as structures that aid goal setting, feedback, and the distribution of rewards contingent on goal completion. Drawing inspiration from how traditional goal setting facilitates accountability between learners and a human tutor, the research explores how intelligent systems can similarly scaffold goal and effort regulation processes while preserving student agency in setting goals and learning associated metacognitive skills of effort regulation, calibration, and goal selection.
Specifically, this work presents the design, implementation, and evaluation of intelligent goal-setting support systems embedded within active learning platforms. Theoretical predictions center around estimating the utility of data-driven goal recommendations, performance feedback, and goal tracking mechanisms using learning system log data. Experimental studies in K-12 classroom settings, centered in grades 5-9, evaluate the effectiveness of adaptive goal-setting interventions on student engagement and learning outcomes. A secondary focus is placed on modeling how achievement events and goal adjustment moments influence student trajectories over time, investigating whether adaptive support can propel students into positive achievement cycles that may foster independent effort regulation.
Empirical findings demonstrate significant engagement and skill mastery benefits of goal-setting support in classrooms beyond tutoring without goal setting. Goal achievement was linked to larger intervention benefits, and adaptive goals, as well as data-driven feedback to guide goal adjustments, were found to enhance student goal completion rates significantly. Though broadly effective for over 80% of students, students who benefited less demonstrated higher levels of baseline engagement. It is conjectured that these students may already exhibit high degrees of intrinsic motivation, which are known to offset reward benefits for performance when rewards are clearly tied to performance or goals are perceived as coercive.
To test this account, the final proposed study will investigate the at-scale application of more frequent goal adjustments delivered through dashboards that enhance frequent student goal choice without researcher facilitation on-site. Explanatory variables assessed through surveys of intrinsic motivation, perceived choice, as well as goal orientation will further explicate and test this interpretation. Findings from this study are expected to guide future efforts in psychometric assessment that could guide finer-grained adaptivity in effort goal delivery through data.
Findings of this thesis advance SRL theories in adaptive learning systems by adding an adaptivity loop to the regulation of effort outside of practice, in addition to metacognitive support during learning. Practically, this work provides empirical evidence on the efficacy of adaptive goal recommendations and offers a scalable solution that alleviates teacher workload. Finally, this research advances scientific understanding of the longitudinal effects of incentivizing effort regulation through extrinsic rewards across different learners, finding that beneficial goal-setting effects are generally sustained over time, but also can lead to differential effects if learners lack goal choice or scaffolds to set realistic goals that they can achieve. These insights can help make educational technologies for deliberate practice in STEM more effective for knowledge acquisition.
Public Link to Thesis Proposal Document:
https://cborchers.com/pdf/dissertation-proposal-draft.pdf