Publications
Building on previous work to understand mental transformation of solid objects, our work aims to contribute to theory on how individuals visualize fluid transformation. Working at the intersection of discipline-based research in fluid-Earth sciences (oceanography and atmospheric science) and cognitive science, we began by characterizing students' mental models of fluids in solid body rotation. Using a common pedagogical tool of oceanography and atmospheric science instructors—large water filled rotating tanks—we conducted semi-structured, interactive interviews in conjunction with demonstrations of solid body fluid rotation to elicit participants' sketched predictions and explanations of their reasoning. We used emergent coding to analyze the data and found that participants confidently predicted fluid behavior; however, their predictions were highly inaccurate, much to the participants' surprise. These results suggest that the behavior of rotating fluids is unintuitive, and that the mind adopts prior experiences with fluids to fill the gap. We suggest that a better understanding of how humans reason about fluids can inform the cognitive science of event processing and suggest that framing mental transformation using an event perspective may be a productive approach.
This project contains the data files associated with the analyses presented in the research article entitled "What drives change? Characterizing scientific self-efficacy development in undergraduate research experiences." (2026-02-02)
Many problems seem to require a flash of insight to solve. What form do these sudden insights take, and what impact do they have on how people approach similar problems in the future? In this work, we prompted participants (N = 189) to think aloud as they attempted to solve a sequence of five "matchstick-arithmetic" problems. These problems either all relied on the same kind of non-obvious solution (Same group) or a different kind each time (Different group). Our first observation was that Same participants improved more rapidly than Different participants. We then leveraged techniques from natural language processing to analyze participants' speech, and found that this accelerated improvement for Same participants was accompanied by changes in both how much they spoke and what they said. In particular, they were more likely to spontaneously label the kind of problem they were working on. Taken together, these findings suggest that a hallmark of transferable insights is their accessibility for verbal report, even if the underlying precursors of insight remain difficult to articulate.
ABSTRACT Background Video data provides rich opportunities to examine student behaviour in game‐based learning environments, capturing not only observable actions but also subtle indicators of cognitive engagement. However, traditional video analysis is labor‐intensive, and current applications of AI to this task are limited by models trained on general‐purpose datasets, which often fail to capture the pedagogical meaning of student actions in authentic educational contexts. Objectives This study explores how multimodal large language models (MLLMs), specifically LLaVA‐Video‐7B‐Qwen2, can support qualitative video analysis of student exploration behaviours in a Minecraft‐based STEM learning environment. Methods We conducted an exploratory case study using screen recordings from a Minecraft‐based STEM learning environment. We tested multiple prompt strategies for guiding MLLM‐generated video descriptions and found that role‐assignment prompting performed most effectively. We evaluated model outputs using a mixed‐method framework that included quantitative scoring, GPT‐based judgement, and researcher validation. Results and Conclusions Our findings show that MLLMs can reliably identify surface‐level behaviours, such as navigation patterns and object interactions, but struggle to infer the intent or goals behind student actions, leading to a significant rate of over‐interpretation (26.5% when explaining student strategies). The model's outputs are sensitive to prompt phrasing, underscoring the importance of prompt engineering. While current MLLMs show promise for streamlining parts of the video analysis workflow, their use in educational contexts requires structured oversight and careful interpretation to ensure reliability and relevance.
Family storytelling can support the development of children's cognitive skills and sense of identity. The current study explored Latine caregivers' storytelling as a resource for promoting their young children's science learning. One-hundred and fifteen U.S. Latine caregivers were video recorded sharing a nature- or science-related story with their 3- to 5-year-old children and reporting why they shared that story. Most caregivers told autobiographical narratives, including reminiscing stories about shared events and caregiver memories from their own childhood. Others told fiction stories (i.e. invented stories, retellings of popular children's stories) and fictionalized personal stories. Stories covered many science topics, including animals, outdoor recreation, natural environments, and plants and farming. Although STEM learning is often presented as acultural and objective, many caregivers chose to tell nature- and science-related stories that connected to child and family identity development.
This dissertation studies how artificial intelligence, especially large language models such as GPT, can help students learn introductory computer programming when the models are used inside a carefully designed learning system. Instead of focusing on AI as a standalone tool, the dissertation follows a connected story: generating learning resources, building a tutoring platform, running a user study, and then analyzing how students behave while they program. The work first uses prompt engineering to create a large collection of 11,700 Python exercises aligned with introductory computer science topics. Students and instructors then evaluate these exercises to check whether they are clear, relevant, and useful for learning. The dissertation next introduces CodeMasterAI, an intelligent tutoring system that presents these exercises in an interactive coding environment and provides AI-generated feedback without simply giving away the answer. Using CodeMasterAI, the dissertation conducts a user study with 19 undergraduate students working on nine programming exercises. The platform records keystrokes, code snapshots, feedback requests, and survey responses, which makes it possible to study learning as it happens rather than only looking at final submissions. The later chapters use these data to examine cognitive load, how students use AI feedback, and how different learner profiles can be detected from programming behavior. Overall, this work shows that large language models can be valuable tools for introductory programming education when they are paired with human evaluation, thoughtful system design, and detailed learning analytics. The dissertation provides both practical guidance for building AI-supported tutoring tools and research evidence for future personalized learning systems.
Recent research has found that the cumulative amount of time a child has had auditory access to language accounts for differences in early numerical abilities between oral Deaf and Hard-of-Hearing (DHH) and hearing preschoolers. We replicate and extend these findings by exploring the role of language ability and salience of numerical information with a larger sample (N = 89; 43 female; majority White) of 3- 6-year-old oral DHH and hearing children. We captured number knowledge, numerical discrimination, salience of numerical information, and parent-reports of vocabulary and language experience. Oral DHH children underperformed on the numerical tasks and vocabulary, however, disparities in numerical measures disappeared when controlling for differences in language experience. The salience of numerical information was similar for both groups.
This paper offers a critical reflection of one intergenerational team's journey to collaborate on a longitudinal study of youth pathways where six participants in the study worked as co-researchers alongside established researchers. By engaging in a careful consideration of our co-research process over the course of multiple years, we articulate our individual and collective experiences of trying, and sometimes failing, to engage in an anti-adultist approach to participatory research. We discuss inherent tensions that emerge when participants who serve as co-researchers and established researchers navigate power dynamics regarding trust and expertise and when traditional research practices nudge out the knowledge and voices of young co-researchers.
Investing in College Readiness: Societal Benefits and Costs of the El Dorado College Promise Program
A growing volume of research shows college promise programs increase the likelihood that students enroll in and complete college. Place-based promise programs provide a guaranteed college scholarship for students who attend a specified school district. We examine the societal benefits and costs of a place-based scholarship program in rural Arkansas, the El Dorado Promise. Our cost framework treats tuition payments as a cash transfer that shifts the cost burden of higher education rather than causing new societal costs. Findings show the program provides societal benefits equal to $4.60 for each dollar invested. A break-even analysis shows the program produces benefits to society that exceed costs within an 81% confidence interval of effects, providing relatively high likelihood of positive returns.
Computers and geospatial technologies such as Geographic Information Systems, Google Earth, and the Global Positioning System (GPS) are increasingly common tools for geography instruction in schools. The NAEP Data Explorer was used to conduct an analysis of geography achievement in the eighth grade based on differences in students' access to and uses of computers and other technologies at home and school. The analysis over the study period (1994-2018) documents the emergence of technologies available to support geography instruction. Results point to differences in geography achievement associated with differential access to technology and applications of technology for geographic learning.


