VIABLE Lab Presents at LAK 2026 in Bergen

The VIABLE Lab recently participated in the 16th International Learning Analytics & Knowledge Conference (LAK26), held from April 27-May 1, 2026 in Bergen, Norway. Shan Zhang represented the lab on site, presenting two studies that reflect our continuing work in learning analytics, AI literacy, mathematics learning, and educational measurement.

LAK is one of the central venues for research on how data, models, and analytics can help us understand learning more carefully. This year's conference offered a strong setting for conversations about what can be inferred from learner behavior, how educational constructs should be measured, and how analytics can support better decisions without oversimplifying the people behind the data.

Conference Participation

Through LAK 2026, we had the opportunity to share research findings with an international community of researchers, practitioners, and industry partners working in learning analytics, educational data mining, artificial intelligence in education, and the learning sciences. Our presentations approached measurement from different directions: one examined students' replay behavior in a game-based mathematics learning environment, while the other examined how K-12 teachers' self-reported AI literacy aligns with objective-based measures. Together, they point to a shared message: educational data can be powerful, but only when interpreted with attention to timing, context, and the limits of any single measure.

Research Presentations

Replay Behavior and Learning Pathways

The study "Let Me Try Again: Examining Replay Behavior by Tracing Students' Latent Problem-Solving Pathways" explored how students revisit previously completed math problems and whether replay behavior supports mathematics learning. Using latent pathway analysis, the research found that replay is not always equally beneficial. Immediate replay was associated with stronger learning outcomes, including conceptual understanding, procedural knowledge, and mathematical flexibility, whereas delayed replay showed weak or even negative associations with learning outcomes. These findings suggest that when students choose to revisit learning activities may be just as important as whether they revisit them at all, providing new insights for the design of adaptive educational technologies and learning platforms.

Assessing AI Literacy Among Teachers

The study "How to Assess AI Literacy: Misalignment Between Self-Reported and Objective-Based Measures" examined both self-reported and objective-based AI literacy assessments among K-12 teachers using a unified AI literacy framework. Results revealed only weak alignment between what teachers believed they knew about AI and what they could objectively demonstrate. The analysis further identified distinct teacher profiles, including overestimators and underestimators, highlighting the importance of using multiple assessment approaches when evaluating AI literacy. This work contributes to ongoing efforts to develop more valid and comprehensive measures of AI literacy for educators.

What These Studies Add

Although the two papers focus on different domains, they both caution against easy interpretations. Replaying a problem is not automatically productive just because it happens. Feeling confident about AI is not the same as demonstrating AI literacy. In both cases, better measurement helps researchers, designers, and educators see the learning situation more clearly.

That perspective is especially important as educational systems increasingly rely on analytics, AI tools, and dashboards. The goal is not simply to gather more data, but to build interpretations that remain accountable to students, teachers, and the complexity of learning itself.

Presented Work

[1] Zhang, S., Pradhan, S., Lee, J.-E., Gurung, A., & Botelho, A. F. (2026). Let Me Try Again: Examining Replay Behavior by Tracing Students' Latent Problem-Solving Pathways. In Proceedings of the 16th International Learning Analytics & Knowledge Conference (LAK 2026), Bergen, Norway, 425-435. ACM Digital Library

[2] Zhang, S., Xiao, R., Botelho, A. F., Liao, G., Chiu, T. K. F., Stamper, J., & Koedinger, K. R. (2026). How to Assess AI Literacy: Misalignment Between Self-Reported and Objective-Based Measures. In Proceedings of the 16th International Learning Analytics & Knowledge Conference (LAK 2026), Bergen, Norway, 405-414. ACM Digital Library

Bergen Between Sessions

Beyond the conference, Bergen offered a memorable experience of its own: colorful waterfront buildings around Bryggen, mountain views above the harbor, and the sense of fjords never far from the city. Exploring the historic district and riding the Fløibanen funicular toward Fløyen made the academic week feel rooted in a very specific place.

The long Nordic summer evenings helped stretch the days in a good way. After sessions on replay behavior, AI literacy, and learning analytics, there was still light left for walking, looking back over the harbor, and letting the city reset the pace before the next round of conference conversations.