Xinying Hou

There are growing efforts to increase the percentage of secondary students who take computing courses. However, U.S. women identifying as Black/African American, Hispanic/Latina, and/or Native American remain underrepresented in computing and face persistent challenges in learning to program. Moreover, it remains underexplored what types of programming learning support minoritized high school women desire. To address this gap, this study investigates the challenges minoritized high school women encounter in current helpseeking practices and explores the forms of programming learning support they desire. We conducted design sessions with 12 high school women who are minoritized in computing and enrolled in Advanced Placement Computer Science courses. Using thematic analysis, we examined the challenges they encountered with existing help-seeking resources, as well as their preferred format and support features.

As AI enters computer science (CS) education, novices need early opportunities to strengthen code understanding. This work responds to this need by investigating pairing mixed-up code puzzles (Parsons puzzles) with self-explanation (SE) prompts. Drawing on learning sciences work, we designed two SE prompts: Select, where students engage in SE by choosing correct options, and Fill, where they engage in SE by typing answers. In an in-depth within-subjects study (N=10), all novices valued SE prompts for understanding code after solving puzzles. Select was strongly preferred to Fill for its efficiency and scaffolding. Fill was removed due to frustration from likely non-learning mental effort. A between-subjects study (N=110) in an intro-level CS classroom compared practice with puzzles-only to puzzles plus Select SEs. The group solving puzzles with Select showed significantly greater pre-to- posttest learning gains with similar perceived in-practice mental effort.

As AI enters computer science (CS) education, novices need early opportunities to strengthen code understanding. This work responds to this need by investigating pairing mixed-up code puzzles (Parsons puzzles) with self-explanation (SE) prompts. Drawing on learning sciences work, we designed two SE prompts: Select, where students engage in SE by choosing correct options, and Fill, where they engage in SE by typing answers. In an in-depth within-subjects study (N=10), all novices valued SE prompts for understanding code after solving puzzles. Select was strongly preferred to Fill for its efficiency and scaffolding. Fill was removed due to frustration from likely non-learning mental effort. A between-subjects study (N=110) in an intro-level CS classroom compared practice with puzzles-only to puzzles plus Select SEs. The group solving puzzles with Select showed significantly greater pre-to- posttest learning gains with similar perceived in-practice mental effort.

Generative artificial intelligence (GenAI) is rapidly entering K-12 classrooms worldwide, initiating urgent debates about its potential to either reduce or exacerbate educational inequalities. Drawing on interviews with 30 K-12 teachers across the United States, South Africa, and Taiwan, this study examines how teachers navigate this GenAI tension around educational equalities. We found teachers actively framed GenAI education as an equality-oriented practice: they used it to alleviate pre-existing inequalities while simultaneously working to prevent new inequalities from emerging. Despite these efforts, teachers confronted persistent systemic barriers, i.e., unequal infrastructure, insufficient professional training, and restrictive social norms, that individual initiative alone could not overcome. Teachers thus articulated normative visions for more inclusive GenAI education. By centering teachers’ practices, constraints, and future envisions, this study contributes a global account of how GenAI education is being integrated into K-12 contexts and highlights what is required to make its adoption genuinely equal.

Conversation-based assessments (CBA), which evaluate student knowledge through interactive dialogues with artificial agents on a given topic, can help address non-effortful formative test-taking and the lack of adaptability in traditional assessment. This work employs evidence-centered design framework with LLM techniques to establish a multi-agent architecture for conversation-based assessment. It includes four LLM agents: two student-facing agents and two behind-the-scenes agents. All agents are monitored by a non-LLM agent (Watcher), which manages the assessment flow through updated instructions to agents and turn control.

With recent developments, GenAI applications have begun supporting multiple modes of communication, known as multimodality. In this work, we explored how undergraduate programming novices choose and work with multimodal GenAI tools, and their criteria for choices. We selected a commercially available multimodal GenAI platform for interaction, as it supports multiple input and output modalities, including text, audio, image upload, and real-time screen-sharing. Through 16 think-aloud sessions that combined participant observation with follow-up semi-structured interviews, we investigated student modality choices for GenAI tools when completing programming problems and the underlying criteria for modality selections.

We first proposed pedagogical prompting, a theoretically-grounded new concept to elicit learning-oriented responses from LLMs. For proof-of-concept learning intervention in a real educational setting, we selected early undergraduate CS education (CS1/CS2) as the example context. Based on instructor insights, we designed and developed a learning intervention as an interactive system with scenario-based instruction to train pedagogical prompting skills. Finally, we assessed its effectiveness with pre/post-tests in a user study of CS undergraduates.

We presented CodeTailor, a system that leverages a large language model (LLM) to provide personalized help to students while still encouraging cognitive engagement. CodeTailor provides a personalized Parsons puzzle to support struggling students. In a Parsons puzzle, students place mixed-up code blocks in the correct order to solve a problem. CodeTailor distinguishes itself from existing LLM-based products by providing an active learning opportunity where students are expected to "solve" the puzzle rather than simply acting as passive consumers by "read- ing" a direct solution.

We presented the results of a thematic analysis on a data set from 33 learners as they independently learned Python by working on 45 code-authoring tasks with access to an AI Code Generator based on OpenAI Codex. Our analysis reveals four distinct coding approaches when writing code with an AI code generator: AI Single Prompt; AI Step-by-Step; Hybrid; and Manual coding, where learners wrote the code themselves.

In this paper, we explore using Parsons problems to scaffold novice programmers who are struggling while solving write-code problems. Parsons problems, in which students put mixed-up code blocks in order, can be created quickly and already serve thousands of students while other types of programming support methods are expensive to develop or do not scale. We conducted studies in which novices were given equivalent Parsons problems as optional scaffolding while solving write-code problems. We investigated when, why, and how students used the Parsons problems as well as their perceptions of the benefits and challenges.

In this math digital learning game, one version encouraged playing and learning through an open learner model, while one encouraged playing for enjoyment through an analogous open enjoyment model. We compared these versions to a control version that is neutral with respect to learning and enjoyment. The learning-oriented group engaged more in re-practicing, while the enjoyment-oriented group demonstrated more exploration of different mini-games. In turn, our analyses have led to preliminary ideas about how to use AI to provide recommendations that are more aligned with students’ dynamic learning and enjoyment states and preferences.