Ongoing Research
1. AI-Assisted Learning and Metacognitive Calibration
Research focus: Metacognition, AI-assisted reasoning, and cognitive independence
Does access to generative AI improve students' reasoning, or can it create a mismatch between their actual performance and their confidence in what they know?
This experimental study investigates how AI assistance influences students' performance, confidence judgments, and metacognitive accuracy during logical reasoning tasks. We examine whether AI affects their ability to accurately evaluate their own performance and identify individual differences that may influence these effects.
The project contributes to a broader understanding of how AI–learner interactions can augment cognitive performance while potentially undermining the metacognitive processes essential for independent learning.
2. Generative AI, Study Strategies, and Academic Performance
Research focus: AI use, self-regulated learning, and exam performance
How students use AI may matter more for learning than whether they use it at all.
Our research examines how undergraduate students integrate generative AI into their exam preparation, which study strategies they use alongside AI, and how these practices relate to academic performance.
The initial study, conducted in a large undergraduate cognitive psychology course, examined students' self-reported AI use and study strategies across multiple exams. Findings suggest that AI use alone does not explain differences in academic performance. Instead, the ways students engage with AI, the learning strategies they employ, and their participation in course activities provide a more informative picture of academic outcomes.
Current expansion: The project is being extended to undergraduate mathematics courses, allowing us to investigate whether these patterns generalize across disciplines and different types of academic tasks.
3. Evaluating the Learning Hub Intervention
Research focus: Cognitive science-based learning strategies, metacognition, and academic success
Can explicit instruction in evidence-based learning strategies improve students' academic skills and support lasting changes in study behavior?
This project evaluates the Learning Hub's six-session intervention designed to help first-year students develop more effective learning strategies. Using pre- and post-intervention assessments, comparison groups, and follow-up measures, we investigate changes in metacognitive self-regulation, critical thinking, study organization, time management, academic self-efficacy, and academic outcomes.
Evaluations conducted in Fall 2024 and Fall 2025 have provided encouraging evidence of improvements in several learning strategy domains, particularly among students participating in mentor-led, in-person sessions.
Ongoing research examines the persistence of these changes, their transfer to academic coursework, and opportunities to improve the intervention.
4. Human- and AI-avatar-guided Learning Interventions
Research focus: AI-supported instruction, learning intervention design, and educational accessibility
Can AI-avatar-guided learning experiences provide effective instruction in evidence-based study strategies, and how do their outcomes compare with human mentorship?
Building on the Learning Hub's in-person intervention, this research investigates the potential of AI-guided online learning modules to expand access to cognitive science-based instruction. Our initial evaluation compared in-person instruction, online learning modules, and a control condition. Results identified different strengths across instructional formats, informing the next stage of intervention development.
Current work focuses on refining the online learning experience, evaluating its usability, and developing approaches that preserve meaningful cognitive engagement while incorporating AI-based instructional support.
This research also informs a proposed larger-scale evaluation comparing mentor-led, AI-guided, and control conditions.
5. AI-Generated Educational Assessments
Research focus: Artificial intelligence, psychological measurement, and educational assessment
How can generative AI support the development of valid and reliable measures of students' academic skills, attitudes, and learning-related characteristics?
Through collaborations in AI-based educational assessment, we investigate methods for generating and validating assessment items using large language models and modern psychometric approaches. The Learning Hub provides opportunities to study these methods with undergraduate learners and examine how AI-generated assessments correspond to students' actual responses.
This work supports the broader goal of developing rigorous, scalable tools for evaluating educational interventions and understanding individual differences in learning.
Completed Research
6. Using AI to Understand Student Feedback and Improve Teaching
Research focus: AI-based text analysis, student engagement, and active learning
Large university courses generate substantial amounts of student feedback, but traditional course evaluations often fail to capture the full complexity of students' learning experiences. This project investigated how AI-based natural language processing can help instructors analyze open-ended student feedback and identify meaningful patterns in engagement, learning, and classroom experiences. Applying these methods to large undergraduate psychology courses, the research identified differences in students' reported learning experiences across traditional and active-learning instructional formats. The findings demonstrate how AI can support educational research and provide instructors with actionable insights for improving teaching.
Related publications:
- Teles, M., & Tomasevic, A. (2026). Making sense of students' feedback in large classes: A practical guide using AI-based text analysis. College Teaching.
- Teles, M., Clabough, E., & Byron, T. (2026). Transforming a large-enrollment neuroscience course with active learning: Using AI-based text analysis to capture what traditional evaluations miss. Journal of Undergraduate Neuroscience Education.
Foundations in Cognitive Science Research
Cognitive Aging, Memory, and Cognitive Interventions
The Learning Hub builds on a broader research foundation in human memory, cognitive development, and evidence-based cognitive interventions.
Previous research conducted through the NIH-funded Virginia Cognitive Aging Project and related studies has investigated memory processes, cognitive change across adulthood, and factors influencing learning and retrieval. These research experiences inform our current interest in understanding how cognitive processes can be strengthened through targeted interventions and how new technologies may influence human learning and cognition.
Undergraduate Research Opportunities
Undergraduate students play an essential role in the Learning Hub's research activities.
Through research assistantships and experiential research methods courses, students participate in study design, data collection, statistical analysis, interpretation of findings, and research presentations.
Our goal is to provide meaningful opportunities for students to engage in the scientific process while contributing to research that advances our understanding of learning.
Interested in joining our research team?
Contact Dr. Mariana Teles at mt2yq@virginia.edu.