Companion Proceedings of the 15th International Learning Analytics and Knowledge Conference (LAK’25)

We are very pleased to welcome you to the Fifteenth International Conference on Learning Analytics and Knowledge (LAK25), organized by the Society for Learning Analytics Research (SoLAR). This year’s conference is held in Dublin, Ireland, between March 3rd and 7th, 2025. 

The theme for the 15th annual LAK conference is Expanding the Horizons of Learning Analytics. After this many years of research and practice, the learning analytics field has established its own identity, traditions, and community. Pursuing our initial objective of making use of data to better understand and improve learning processes, we have studied and impacted numerous aspects of both formal and informal education. However, as the field enters its teenage years, it faces swift and significant shifts in technological, theoretical, and pedagogical contexts that have a direct effect on our work. For instance, artificial intelligence offers the yet-to-be-proven promise of facilitating and democratizing data analysis, while also posing significant ethical challenges. Critical theories prompt us to examine the values and unintended consequences of our contributions. Novel educational models demand innovative methods for studying learning processes and measuring and assessing learning outcomes. In response, the community is actively reassessing and reshaping what it means to “do learning analytics” within these evolving environments. This process of reinvention often involves stepping out of our comfort zone, established during the field’s formative years, to explore new theories, learning processes, data sources, communication modalities, analytical methods, delivery mechanisms, structures for ownership and adoption, and even reconsidering who leads and implements the analytics process. This year’s conference aims to highlight and celebrate the trailblazing works that are expanding the horizons of the learning analytics field.

Two excellent keynote talks and a keynote panel present compelling examples of expanding the horizons of learning analytics, but also raise important questions regarding the effects of such an expansion on the learning analytics field itself. Inge Molenaar is the Director of the National Education Lab AI (NOLAI) and a Professor of Education and Artificial Intelligence at the Behavioural Science Institute, Radboud University, Netherlands. Inge’s keynote explores the role of artificial intelligence (AI) in education, highlighting the dual role of AI – as both a tool and an actor – and emphasizing the potential for hybrid human-AI collaboration. Gautam Biswas is a Cornelius Vanderbilt Professor of Engineering at Vanderbilt University, whose research focuses on developing intelligent, open-ended learning environments for STEM and computer science education. In his keynote, Gautam presents the work that he and his research team have done on the design, development, and deployment of a multimodal, theoretically grounded learning analytics framework, to analyze and interpret students’ collaborative behaviors in STEM environments. The last day of the conference starts with an interactive keynote panel, facilitated by four outstanding learning analytics researchers, namely Rebecca Ferguson (The Open University), Kirsty Kitto (University of Technology Sydney), and Catherine Manly (Fairleigh Dickinson University). Titled “Learning analytics 2035: Pushing the boundaries and meeting the challenges”, this interactive panel invites the learning analytics community to ponder on and discuss some of the grand challenges that have been identified during LAK25, explore ways for addressing them and consider what might happen if these are ignored. The conference features two additional panels. One is focused on opportunities, challenges, and risks of adopting learning analytics in higher education settings, whereas the other re-examines the connections and “boundaries” between learning analytics and closely related research fields that rely on educational data (e.g., educational data mining, learning at scale, and quantitative ethnography), in the light of increasing focus on (Generative) AI across all these fields.

This year’s conference theme encouraged researchers and practitioners to consider distinct ways of extending the horizons of learning analytics such as proposing novel methods and approaches for data collection, analyses, and communication of analytics results, as well as bringing learning analytics to novel or underexplored learning settings and learning processes, and dealing with ethical issues that novel technologies and learning contexts introduce. This encouragement might partially explain a very large number of high-quality submissions we have received this year, breaking all previous records, and we are extremely grateful for all those who decided to submit the results of their latest research efforts to LAK25. The research track had 337 submissions (232 full paper submissions and 105 short paper submissions). This represents an increase of about 7% in the total number of submissions compared to last year. These papers came from research institutions of 28 countries (11 in Europe, 9 in Asia, 2 in Middle East, 2 in South America, 2 in North America, and 1 in Oceania). Maintaining the high quality of the conference, the program committee for the research track consisted of 280 researchers from the field of learning analytics, educational data mining, learning sciences, educational technology, and related disciplines. Of these, 78 were senior members, all recognized leaders in the field and highly involved in service to the learning analytics community. Overall, from the 337 research submissions, the program committee worked very hard to select 101 papers (70 full research papers and 31 short research papers) that are included in the proceedings of the 15th Learning Analytics and Knowledge Conference. The acceptance rate for both full and short research tracks is 30%. 

The rigorous selection process for LAK includes an initial phase of review of at least two program committee members. Authors are then given a short time to provide an optional rebuttal to the remarks and comments raised in the initial review in which they can answer specific questions raised by reviewers (if any) or flag any inaccuracies, omissions, or errors in the reviews. This is followed by the meta-review phase during which, for each submission, a senior program committee member, having carefully reviewed the initial reviews and the authors’ rebuttal (if submitted), provides a summary meta-review and final recommendation to the program chairs. We are most grateful for all the hard work by the program committee and their insightful and constructive comments and reviews. These proceedings could not have been possible without their generous help and support. 

We would also like to emphasize our ongoing gratitude for the efforts made by all involved in the learning analytics community. We very much understand the complexity of work and life pressures impacting on our time commitments, and priorities. The high level of support and commitment shown by our colleagues to ensure that the presented and published papers have received high quality reviews and feedback is highly valued and appreciated. We want to thank you for the important efforts you have devoted that have allowed this conference to continue as a premier scientific event fostering the scholarly exchange of ideas of the highest caliber. 

We hope that LAK25 participants and other readers of these proceedings will find value in the broad range of contributions to the field of learning analytics contained within. The rapid development and adoption of AI-based technologies, especially generative AI, as well as technological developments more broadly are opening many new opportunities for learning analytics research and practice, but also introducing novel challenges that call for novel methodological approaches and theoretical frameworks. Likewise, further work and novel approaches are needed to assure responsible use and analytics of learning-related data, meet the needs and expectations of diverse stakeholders, as well as ensure ethical conduct in learning analytics research and practice and fair and just treatment of all learners. We hope that the scholarly exchanges at this conference, including the paper presentations, keynotes, panels, and both formal and informal discussions among the participants will contribute to addressing the aforementioned and related challenges and bring us closer to the ultimate objective of understanding and advancing learning and the environments in which it occurs. 

LAK25 Companion Proceedings