arXiv subjects

Ryan S. Baker

Publications and source records attributed to Ryan S. Baker.

At least 19 recordsLinked to original sources

arXiv2026

A Case Study Reexamining the Cold-Start Problem in Knowledge Tracing Models and Implications for SafeInsights, an Education Research Infrastructure

Knowledge tracing (KT) models are widely used to predict students' evolving knowledge states from their learning history. However, many KT models are evaluated using specific datasets, platforms, and learning contexts, raising questions about whether reported model performance replicates and generalizes across newer datasets that vary in context. This paper replicates and extends Zhang et al. (2021), which examined the cold-start problem in KT models and found that deep-learning-based KT models performed better, partly because of stronger predictions when students began practicing a skill. Using a more recent ASSISTments dataset, FoundationalASSIST, we replicate the previous analysis by evaluating model performance across opportunities to practice and extend the analysis by examining performance across problem types, including fill-in-the-blank, multiple-choice select-one, multiple-choice select-all, and order/sort problems. Results show that KT model performance varies across both student practice trajectories and problem types. Beyond the empirical replication, this study identifies practical challenges in reproducing educational data mining studies and serves as a proof of concept, showing how privacy-preserving research infrastructures such as SafeInsights can be leveraged to facilitate educational research and support replication analyses.

cs.HC↗

arXiv2026

Building AI Companions that Prioritise Learning over Performance

Large language models (LLMs) are rapidly transforming knowledge work by improving the quality and efficiency of tasks such as writing, coding, and data analysis. However, their growing use in education has exposed a learning-performance paradox: while they can enhance short-term task performance, they may also undermine genuine learning, including cognitive growth, knowledge transfer, and metacognitive development. This paper addresses the question of how artificial intelligence should be designed and used to support learning rather than merely improve immediate outputs. We introduce the concept of AI learning companions, defined as adaptive, pedagogically informed, LLM-powered agents designed for integration into learning environments. We propose a framework for their design built on three interrelated foundations: a pedagogical foundation focused on how students learn with AI, an adaptive foundation focused on how AI learns about students, and a responsible design foundation ensuring systems remain transparent, accountable, inclusive, and secure. The framework is illustrated through five case studies spanning diverse educational contexts, levels, and tool designs, revealing both the promise and current limitations of existing tools. We conclude that there is a necessary shift away from LLMs designed for task-oriented performance, and beyond simply prompting them to act as tutors, toward deliberately developed AI learning companions that are pedagogically sound, adapt to their learners, and foster durable understanding, metacognitive growth, and learner agency.

cs.HC↗

arXiv2026

The Path to Conversational AI Tutors: Integrating Tutoring Best Practices and Targeted Technologies to Produce Scalable AI Agents

The emergence of generative AI has accelerated the development of conversational tutoring systems that interact with students through natural language dialogue. Unlike prior intelligent tutoring systems (ITS), which largely function as adaptive and interactive problem sets with feedback and hints, conversational tutors hold the potential to simulate high-quality human tutoring by engaging with students' thoughts, questions, and misconceptions in real time. While some previous ITS, such as AutoTutor, could respond conversationally, they were expensive to author and lacked a full range of conversational ability. Generative AI has changed the capacity of ITS to engage conversationally. However, realizing the full potential of conversational tutors requires careful consideration of what research on human tutoring and ITS has already established, while also unpacking what new research will be needed. This paper synthesizes tenets of successful human tutoring, lessons learned from legacy ITS, and emerging work on conversational AI tutors. We use a keep, change, center, study framework for guiding the design of conversational tutoring. We argue that systems should keep proven methods from prior ITS, such as knowledge tracing and affect detection; change how tutoring is delivered by leveraging generative AI for dynamic content generation and dialogic scaffolding; and center opportunities for meaning-making, student agency, and granular diagnosis of reasoning. Finally, we identify areas requiring further study, including efficacy testing, student experience, and integration with human instruction. By synthesizing insights from human tutoring, legacy ITS, and emerging generative AI technologies, this paper outlines a research agenda for developing conversational tutors that are scalable, pedagogically effective, and responsive to the social and motivational dimensions of learning.

cs.HC↗

arXiv2026

Understanding Gaming the System by Analyzing Self-Regulated Learning in Think-Aloud Protocols

In digital learning systems, gaming the system refers to occasions when students attempt to succeed in an educational task by systematically taking advantage of system features rather than engaging meaningfully with the content. Often viewed as a form of behavioral disengagement, gaming the system is negatively associated with short- and long-term learning outcomes. However, little research has explored this phenomenon beyond its behavioral representation, leaving questions such as whether students are cognitively disengaged or whether they engage in different self-regulated learning (SRL) strategies when gaming largely unanswered. This study employs a mixed-methods approach to examine students' cognitive engagement and SRL processes during gaming versus non-gaming periods, using utterance length and SRL codes inferred from think-aloud protocols collected while students interacted with an intelligent tutoring system for chemistry. We found that gaming does not simply reflect a lack of cognitive effort; during gaming, students often produced longer utterances, were more likely to engage in processing information and realizing errors, but less likely to engage in planning, and exhibited reactive rather than proactive self-regulatory strategies. These findings provide empirical evidence supporting the interpretation that gaming may represent a maladaptive form of SRL. With this understanding, future work can address gaming and its negative impacts by designing systems that target maladaptive self-regulation to promote better learning.

cs.CY↗

arXiv2025

The Quick Red Fox gets the best Data Driven Classroom Interviews: A manual for an interview app and its associated methodology

Data Driven Classroom Interviews (DDCIs) are an interviewing technique that is facilitated by recent technological developments in the learning analytics community. DDCIs are short, targeted interviews that allow researchers to contextualize students' interactions with a digital learning environment (e.g., intelligent tutoring systems or educational games) while minimizing the amount of time that the researcher interrupts that learning experience, and focusing researcher time on the events they most want to focus on DDCIs are facilitated by a research tool called the Quick Red Fox (QRF)--an open-source server-client Android app that optimizes researcher time by directing interviewers to users that have just displayed an interesting behavior (previously defined by the research team). QRF integrates with existing student modeling technologies (e.g., behavior-sensing, affect-sensing, detection of self-regulated learning) to alert researchers to key moments in a learner's experience. This manual documents the tech while providing training on the processes involved in developing triggers and interview techniques; it also suggests methods of analyses.

cs.HC↗

arXiv2025

Automatic Large Language Models Creation of Interactive Learning Lessons

We explore the automatic generation of interactive, scenario-based lessons designed to train novice human tutors who teach middle school mathematics online. Employing prompt engineering through a Retrieval-Augmented Generation approach with GPT-4o, we developed a system capable of creating structured tutor training lessons. Our study generated lessons in English for three key topics: Encouraging Students' Independence, Encouraging Help-Seeking Behavior, and Turning on Cameras, using a task decomposition prompting strategy that breaks lesson generation into sub-tasks. The generated lessons were evaluated by two human evaluators, who provided both quantitative and qualitative evaluations using a comprehensive rubric informed by lesson design research. Results demonstrate that the task decomposition strategy led to higher-rated lessons compared to single-step generation. Human evaluators identified several strengths in the LLM-generated lessons, including well-structured content and time-saving potential, while also noting limitations such as generic feedback and a lack of clarity in some instructional sections. These findings underscore the potential of hybrid human-AI approaches for generating effective lessons in tutor training.

cs.CY↗

arXiv2025

Starting Seatwork Earlier as a Valid Measure of Student Engagement

Prior work has developed a range of automated measures ("detectors") of student self-regulation and engagement from student log data. These measures have been successfully used to make discoveries about student learning. Here, we extend this line of research to an underexplored aspect of self-regulation: students' decisions about when to start and stop working on learning software during classwork. In the first of two analyses, we build on prior work on session-level measures (e.g., delayed start, early stop) to evaluate their reliability and predictive validity. We compute these measures from year-long log data from Cognitive Tutor for students in grades 8-12 (N = 222). Our findings show that these measures exhibit moderate to high month-to-month reliability (G > .75), comparable to or exceeding gaming-the-system behavior. Additionally, they enhance the prediction of final math scores beyond prior knowledge and gaming-the-system behaviors. The improvement in learning outcome predictions beyond time-on-task suggests they capture a broader motivational state tied to overall learning. The second analysis demonstrates the cross-system generalizability of these measures in i-Ready, where they predict state test scores for grade 7 students (N = 818). By leveraging log data, we introduce system-general naturally embedded measures that complement motivational surveys without extra instrumentation or disruption of instruction time. Our findings demonstrate the potential of session-level logs to mine valid and generalizable measures with broad applications in the predictive modeling of learning outcomes and analysis of learner self-regulation.

cs.CY↗

arXiv2025

Detecting Unsuccessful Students in Cybersecurity Exercises in Two Different Learning Environments

This full paper in the research track evaluates the usage of data logged from cybersecurity exercises in order to predict students who are potentially at risk of performing poorly. Hands-on exercises are essential for learning since they enable students to practice their skills. In cybersecurity, hands-on exercises are often complex and require knowledge of many topics. Therefore, students may miss solutions due to gaps in their knowledge and become frustrated, which impedes their learning. Targeted aid by the instructor helps, but since the instructor's time is limited, efficient ways to detect struggling students are needed. This paper develops automated tools to predict when a student is having difficulty. We formed a dataset with the actions of 313 students from two countries and two learning environments: KYPO CRP and EDURange. These data are used in machine learning algorithms to predict the success of students in exercises deployed in these environments. After extracting features from the data, we trained and cross-validated eight classifiers for predicting the exercise outcome and evaluated their predictive power. The contribution of this paper is comparing two approaches to feature engineering, modeling, and classification performance on data from two learning environments. Using the features from either learning environment, we were able to detect and distinguish between successful and struggling students. A decision tree classifier achieved the highest balanced accuracy and sensitivity with data from both learning environments. The results show that activity data from cybersecurity exercises are suitable for predicting student success. In a potential application, such models can aid instructors in detecting struggling students and providing targeted help. We publish data and code for building these models so that others can adopt or adapt them.

cs.LG↗

arXiv2025

Carelessness Detection using Performance Factor Analysis: A New Operationalization with Unexpectedly Different Relationship to Learning

Detection of carelessness in digital learning platforms has relied on the contextual slip model, which leverages conditional probability and Bayesian Knowledge Tracing (BKT) to identify careless errors, where students make mistakes despite having the knowledge. However, this model cannot effectively assess carelessness in questions tagged with multiple skills due to the use of conditional probability. This limitation narrows the scope within which the model can be applied. Thus, we propose a novel model, the Beyond Knowledge Feature Carelessness (BKFC) model. The model detects careless errors using performance factor analysis (PFA) and behavioral features distilled from log data, controlling for knowledge when detecting carelessness. We applied the BKFC to detect carelessness in data from middle school students playing a learning game on decimal numbers and operations. We conducted analyses comparing the careless errors detected using contextual slip to the BKFC model. Unexpectedly, careless errors identified by these two approaches did not align. We found students' post-test performance was (corresponding to past results) positively associated with the carelessness detected using the contextual slip model, while negatively associated with the carelessness detected using the BKFC model. These results highlight the complexity of carelessness and underline a broader challenge in operationalizing carelessness and careless errors.

cs.CY↗

arXiv2024

Exploring Knowledge Tracing in Tutor-Student Dialogues using LLMs

Recent advances in large language models (LLMs) have led to the development of artificial intelligence (AI)-powered tutoring chatbots, showing promise in providing broad access to high-quality personalized education. Existing works have studied how to make LLMs follow tutoring principles, but have not studied broader uses of LLMs for supporting tutoring. Up until now, tracing student knowledge and analyzing misconceptions has been difficult and time-consuming to implement for open-ended dialogue tutoring. In this work, we investigate whether LLMs can be supportive of this task: we first use LLM prompting methods to identify the knowledge components/skills involved in each dialogue turn, i.e., a tutor utterance posing a task or a student utterance that responds to it. We also evaluate whether the student responds correctly to the tutor and verify the LLM's accuracy using human expert annotations. We then apply a range of knowledge tracing (KT) methods on the resulting labeled data to track student knowledge levels over an entire dialogue. We conduct experiments on two tutoring dialogue datasets, and show that a novel yet simple LLM-based method, LLMKT, significantly outperforms existing KT methods in predicting student response correctness in dialogues. We perform extensive qualitative analyses to highlight the challenges in dialogueKT and outline multiple avenues for future work.

cs.CL↗

arXiv2024

ABROCA Distributions For Algorithmic Bias Assessment: Considerations Around Interpretation

Algorithmic bias continues to be a key concern of learning analytics. We study the statistical properties of the Absolute Between-ROC Area (ABROCA) metric. This fairness measure quantifies group-level differences in classifier performance through the absolute difference in ROC curves. ABROCA is particularly useful for detecting nuanced performance differences even when overall Area Under the ROC Curve (AUC) values are similar. We sample ABROCA under various conditions, including varying AUC differences and class distributions. We find that ABROCA distributions exhibit high skewness dependent on sample sizes, AUC differences, and class imbalance. When assessing whether a classifier is biased, this skewness inflates ABROCA values by chance, even when data is drawn (by simulation) from populations with equivalent ROC curves. These findings suggest that ABROCA requires careful interpretation given its distributional properties, especially when used to assess the degree of bias and when classes are imbalanced.

stat.ML↗

arXiv2024

Evaluating Algorithmic Bias in Models for Predicting Academic Performance of Filipino Students

Algorithmic bias is a major issue in machine learning models in educational contexts. However, it has not yet been studied thoroughly in Asian learning contexts, and only limited work has considered algorithmic bias based on regional (sub-national) background. As a step towards addressing this gap, this paper examines the population of 5,986 students at a large university in the Philippines, investigating algorithmic bias based on students' regional background. The university used the Canvas learning management system (LMS) in its online courses across a broad range of domains. Over the period of three semesters, we collected 48.7 million log records of the students' activity in Canvas. We used these logs to train binary classification models that predict student grades from the LMS activity. The best-performing model reached AUC of 0.75 and weighted F1-score of 0.79. Subsequently, we examined the data for bias based on students' region. Evaluation using three metrics: AUC, weighted F1-score, and MADD showed consistent results across all demographic groups. Thus, no unfairness was observed against a particular student group in the grade predictions.

cs.LG↗

arXiv2024

Cultural Bias and Cultural Alignment of Large Language Models

Culture fundamentally shapes people's reasoning, behavior, and communication. As people increasingly use generative artificial intelligence (AI) to expedite and automate personal and professional tasks, cultural values embedded in AI models may bias people's authentic expression and contribute to the dominance of certain cultures. We conduct a disaggregated evaluation of cultural bias for five widely used large language models (OpenAI's GPT-4o/4-turbo/4/3.5-turbo/3) by comparing the models' responses to nationally representative survey data. All models exhibit cultural values resembling English-speaking and Protestant European countries. We test cultural prompting as a control strategy to increase cultural alignment for each country/territory. For recent models (GPT-4, 4-turbo, 4o), this improves the cultural alignment of the models' output for 71-81% of countries and territories. We suggest using cultural prompting and ongoing evaluation to reduce cultural bias in the output of generative AI.

cs.CL↗

arXiv2024

On Fixing the Right Problems in Predictive Analytics: AUC Is Not the Problem

Recently, ACM FAccT published an article by Kwegyir-Aggrey and colleagues (2023), critiquing the use of AUC ROC in predictive analytics in several domains. In this article, we offer a critique of that article. Specifically, we highlight technical inaccuracies in that paper's comparison of metrics, mis-specification of the interpretation and goals of AUC ROC, the article's use of the accuracy metric as a gold standard for comparison to AUC ROC, and the article's application of critiques solely to AUC ROC for concerns that would apply to the use of any metric. We conclude with a re-framing of the very valid concerns raised in that article, and discuss how the use of AUC ROC can remain a valid and appropriate practice in a well-informed predictive analytics approach taking those concerns into account. We conclude by discussing the combined use of multiple metrics, including machine learning bias metrics, and AUC ROC's place in such an approach. Like broccoli, AUC ROC is healthy, but also like broccoli, researchers and practitioners in our field shouldn't eat a diet of only AUC ROC.

cs.LG↗

arXiv2024

Comparison of Three Programming Error Measures for Explaining Variability in CS1 Grades

Programming courses can be challenging for first year university students, especially for those without prior coding experience. Students initially struggle with code syntax, but as more advanced topics are introduced across a semester, the difficulty in learning to program shifts to learning computational thinking (e.g., debugging strategies). This study examined the relationships between students' rate of programming errors and their grades on two exams. Using an online integrated development environment, data were collected from 280 students in a Java programming course. The course had two parts. The first focused on introductory procedural programming and culminated with exam 1, while the second part covered more complex topics and object-oriented programming and ended with exam 2. To measure students' programming abilities, 51095 code snapshots were collected from students while they completed assignments that were autograded based on unit tests. Compiler and runtime errors were extracted from the snapshots, and three measures -- Error Count, Error Quotient and Repeated Error Density -- were explored to identify the best measure explaining variability in exam grades. Models utilizing Error Quotient outperformed the models using the other two measures, in terms of the explained variability in grades and Bayesian Information Criterion. Compiler errors were significant predictors of exam 1 grades but not exam 2 grades; only runtime errors significantly predicted exam 2 grades. The findings indicate that leveraging Error Quotient with multiple error types (compiler and runtime) may be a better measure of students' introductory programming abilities, though still not explaining most of the observed variability.

cs.PL↗

arXiv2024

Navigating Compiler Errors with AI Assistance -- A Study of GPT Hints in an Introductory Programming Course

We examined the efficacy of AI-assisted learning in an introductory programming course at the university level by using a GPT-4 model to generate personalized hints for compiler errors within a platform for automated assessment of programming assignments. The control group had no access to GPT hints. In the experimental condition GPT hints were provided when a compiler error was detected, for the first half of the problems in each module. For the latter half of the module, hints were disabled. Students highly rated the usefulness of GPT hints. In affect surveys, the experimental group reported significantly higher levels of focus and lower levels of confrustion (confusion and frustration) than the control group. For the six most commonly occurring error types we observed mixed results in terms of performance when access to GPT hints was enabled for the experimental group. However, in the absence of GPT hints, the experimental group's performance surpassed the control group for five out of the six error types.

cs.SE↗

arXiv2023

Using Think-Aloud Data to Understand Relations between Self-Regulation Cycle Characteristics and Student Performance in Intelligent Tutoring Systems

Numerous studies demonstrate the importance of self-regulation during learning by problem-solving. Recent work in learning analytics has largely examined students' use of SRL concerning overall learning gains. Limited research has related SRL to in-the-moment performance differences among learners. The present study investigates SRL behaviors in relationship to learners' moment-by-moment performance while working with intelligent tutoring systems for stoichiometry chemistry. We demonstrate the feasibility of labeling SRL behaviors based on AI-generated think-aloud transcripts, identifying the presence or absence of four SRL categories (processing information, planning, enacting, and realizing errors) in each utterance. Using the SRL codes, we conducted regression analyses to examine how the use of SRL in terms of presence, frequency, cyclical characteristics, and recency relate to student performance on subsequent steps in multi-step problems. A model considering students' SRL cycle characteristics outperformed a model only using in-the-moment SRL assessment. In line with theoretical predictions, students' actions during earlier, process-heavy stages of SRL cycles exhibited lower moment-by-moment correctness during problem-solving than later SRL cycle stages. We discuss system re-design opportunities to add SRL support during stages of processing and paths forward for using machine learning to speed research depending on the assessment of SRL based on transcription of think-aloud data.

cs.CY↗

arXiv2023

Towards Generalizable Detection of Urgency of Discussion Forum Posts

Students who take an online course, such as a MOOC, use the course's discussion forum to ask questions or reach out to instructors when encountering an issue. However, reading and responding to students' questions is difficult to scale because of the time needed to consider each message. As a result, critical issues may be left unresolved, and students may lose the motivation to continue in the course. To help address this problem, we build predictive models that automatically determine the urgency of each forum post, so that these posts can be brought to instructors' attention. This paper goes beyond previous work by predicting not just a binary decision cut-off but a post's level of urgency on a 7-point scale. First, we train and cross-validate several models on an original data set of 3,503 posts from MOOCs at University of Pennsylvania. Second, to determine the generalizability of our models, we test their performance on a separate, previously published data set of 29,604 posts from MOOCs at Stanford University. While the previous work on post urgency used only one data set, we evaluated the prediction across different data sets and courses. The best-performing model was a support vector regressor trained on the Universal Sentence Encoder embeddings of the posts, achieving an RMSE of 1.1 on the training set and 1.4 on the test set. Understanding the urgency of forum posts enables instructors to focus their time more effectively and, as a result, better support student learning.

cs.LG↗