Search for: All records

The potential of Large Language Models (LLMs) in education is not trivial, but concerns about academic misconduct, misinformation, and overreliance limit their adoption. To address these issues, we introduce MerryQuery, an AI-powered educational assistant using Retrieval-Augmented Generation (RAG), to provide contextually relevant, course-specific responses. MerryQuery features guided dialogues and source citation to ensure trust and improve student learning. Additionally, it enables instructors to monitor student interactions, customize response granularity, and input multimodal materials without compromising data fidelity. By meeting both student and instructor needs, MerryQuery offers a responsible way to integrate LLMs into educational settings. 

Large Language Models (LLMs) have shown promise in educational applications, but challenges such as hallucinations, lack of contextual relevance, and limited personalization impede their practical adoption. To address these issues, my research introduces MerryQuery, an LLM-powered educational agent that integrates Retrieval-Augmented Generation (RAG), rule-based content control, and Reinforcement Learning from Human Feedback (RLHF). The system features a dynamic learning profile module for adaptive personalization and a multi-step verification framework that cross-checks responses against external sources to enhance trustworthiness. A functional prototype of MerryQuery is being piloted in a real-world classroom. Preliminary results demonstrate improved response reliability and student understanding. 

When instructors want to design programming assignments to motivate their students, a common design choice is to have those students write code to make an artifact (e.g. apps, games, music, or images). The goal of this study is to understand the impacts of including artifact creation in a programming assignment on students’ motivation, time on task, and cognitive load. To do so, we conducted a controlled lab study with seventy-three students from an introductory engineering course. The experimental group created a simulation they could interact with – thus having the full experience of artifact creation – while the control group wrote the exact same code, but evaluated it only with test cases. We hypothesized that students who could interact with the simulation they were programming would be more motivated to complete the assignment and report higher intrinsic motivation. However, we found no significant difference in motivation or cognitive load between the groups. Additionally, the experimental group spent more time completing the assignment than the control group. Our results suggest that artifact creation may not be necessary for motivating students in all contexts, and that artifact creation may have other effects such as increased time on task. Additionally, instructors and researchers should consider when, and in what contexts, artifact creation is beneficial and when it may not be 

Many students struggle with decomposition and planning despite the necessity of these skills in computing education. Hence, more tools are needed to scaffold these processes. In this paper, we present Jigsaw, a standalone visual planning tool to help students practice decomposition and planning before writing code. Jigsaw allows students to compose a solution to a new problem based on previously seen “patterns,” such as the accumulator pattern for summing values or the filter pattern for conditional input selection. Students can connect these patterns together to see how data flows between them and define a solution plan. Jigsaw’s goal is to scaffold students’ planning processes by presenting relevant patterns for a given problem. Using a within-subjects design, we evaluated Jigsaw by observing 17 undergraduate students as they planned for and implemented two programming assignments. The experimental task included Jigsaw, and the control task did not. This design aimed to understand how the tool impacted students’ planning and programming process. Subsequently, we conducted interviews with these students regarding their planning and programming experiences with and without Jigsaw. Many students explicitly mentioned they would employ Jigsaw for planning and appreciated the scaffolding it provided. Students also admired the Jigsaw’s novelty in visualizing programming problems. We conclude with our design takeaways and recommendations for future work. 

Background and Context. The importance of CS to 21st-century life and work has made it important to find ways to integrate learning CS and programming into the regular school day. However, learning CS is difficult, so teachers integrating programming need effective strategies to scaffold the learning. In this study, we analyze students’ log data and apply a novel technique to compare Parsons Problems with from-scratch programming in a middle school science class. Objectives. Our research questions aimed to investigate whether, how, and when Parsons Problems improve learning efficiency for a programming exercise within science, utilizing log data analysis and an automated progress detector (SPD). Method. We conducted a study on 199 students in a 6th-grade science course, divided into two groups: one engaged with Parsons problems, and the other, a control group, worked on the same programming task without scaffolding. Then, we analyzed differences in performance and coding characteristics between the groups. We also adopted an innovative application of SPD to gain a better understanding of how and when Parsons problems helped students make more progress on the coding task, with an objective measure of final student grades. Findings. The experimental group, with scaffolding through Parsons Problems, achieved significantly higher grades, spent significantly less time programming, and toggled less between block category tabs. Interestingly, they ran their code more frequently compared to the control group. The SPD analysis revealed that the experimental group made significantly higher progress in all four quartiles of their coding time. Implications. Our findings suggest that Parsons problems can improve learning efficiency by enhancing novices’ learning experience without negatively impacting their performance or grades, which is especially important when programming is integrated into K12 courses. 

In computing classrooms, building an open-ended programming project engages students in the process of designing and implementing an idea of their own choice. An explicit planning process has been shown to help students build more complex and ambitious open-ended projects. However, novices encounter difficulties in exploring and creatively expressing ideas during planning. We present Idea Builder, a storyboarding-based planning system to help novices visually express their ideas. Idea Builder includes three features: 1) storyboards to help students express a variety of ideas that map easily to programming code, 2) animated example mechanics with example actors to help students explore the space of possible ideas supported by the programming environments, and 3) synthesized starter code to help students easily transition from planning to programming. Through two studies with high school coding workshops, we found that students self-reported as feeling creative and feeling easy to communicate ideas; having access to animated example mechanics of an actor help students to build those actors in their plans and projects; and that most students perceived the synthesized starter code from Idea Builder as helpful and time-saving. 

Many students struggle when they are first learning to program. Without help, these students can lose confidence and negatively assess their programming ability, which can ultimately lead to dropouts. However, detecting the exact moment of student struggle is still an open question in computing education. In this work, we conducted a think-aloud study with five high-school students to investigate the automatic detection of progressing and struggling moments using a detector algorithm (SPD). SPD classifies student trace logs into moments of struggle and progress based on their similarity to prior students' correct solutions. We explored the extent to which the SPD-identified moments of struggle aligned with expert-identified moments based on novices' verbalized thoughts and programming actions. Our analysis results suggest that SPD can catch students' struggling and progressing moments with a 72.5% F1-score, but room remains for improvement in detecting struggle. Moreover, we conducted an in-depth examination to discover why discrepancies arose between expert-identified and detector-identified struggle moments. We conclude with recommendations for future data-driven struggle detection systems. 

Many students rely on examples when learning to program, but they often face barriers when incorporating these examples into their own code and learning the concepts they present. As a step towards designing effective example interfaces that can support student learning, we investigate novices' needs and strategies when using examples to write code. We conducted a study with 12 pairs of high school students working on open-ended game design projects, using a system that allows students to browse examples based on their functionality, and to view and copy the example code. We analyzed interviews, screen recordings, and log data, identifying 5 moments when novices request examples, and 4 strategies that arise when students use examples. We synthesize these findings into principles that can inform the design of future example systems to better support students. 

Block-based programming environments, such as Scratch and Snap!, engage users to create programming artifacts such as games and stories, and share them in an online community. Many Snap! users start programming by reusing and modifying an example project, but encounter many barriers when searching and identifying the relevant parts of the program to learn and reuse. We present Pinpoint, a system that helps Snap! programmers understand and reuse an existing program by isolating the code responsible for specific events during program execution. Specifically, a user can record an execution of the program (including user inputs and graphical output), replay the output, and select a specific time interval where the event of interest occurred, to view code that is relevant to this event. We conducted a small-scale user study to compare users’ program comprehension experience with and without Pinpoint, and found suggestive evidence that Pinpoint helps users understand and reuse a complex program more efficiently. 

Our researchers seek to support students in building block-based programming projects that are motivating and engaging as well as valuable practice in learning to code. A difficult part of the programming process is planning. In this research, we explore how novice programmers used a custom-built planning tool, PlanIT, contrasted against how they used storyboarding when planning games. In a three-part study, we engaged novices in planning and programming three games: a maze game, a break-out game, and a mashup of the two. In a set of five case studies, we show how five pairs of students approached the planning and programming of these three games, illustrating that students felt more creative when storyboarding rather than using PlanIT. We end with a discussion on the implications of this work for designing supports for novices to plan open-ended projects.