Background

This workshop is based on the idea that personalisation in learning analytics (LA) is most effective when it is theoretically sound. Although advances in data science have enabled increasingly complex predictive models, these models frequently operate at the population level, overlooking the cognitive, motivational and strategic diversity that characterises real learners (Gašević et al., 2016). As Cronbach (1957) and Cronbach and Snow (1977) argued, individual differences are not mere 'error variance' to be controlled, but central phenomena to be understood. In this spirit, the workshop advocates moving from data-driven optimisation to theory-informed personalisation — approaches that explain why and for whom learning interventions work, rather than merely detecting correlations.

Theory provides the conceptual framework that enables learning analytics to transcend surface-level associations. As Bourdieu famously noted, 'Research without theory is blind, and theory without research is empty' (Bourdieu & Wacquant, 1992). Theories from psychology and the learning sciences, such as self-regulated learning (Panadero, 2017), expectancy–value models, goal orientation theory, cognitive load theory (Kalyuga, 2007) and mindset research, illuminate the mechanisms that shape learners' engagement and performance. These frameworks help to define what ought to be measured, how indicators can be interpreted and how feedback can be meaningfully personalised. Without such grounding, personalisation risks optimising for statistical regularities rather than educationally relevant constructs.

The LAK community has recognised that effective personalisation cannot be ensured by more data or more sophisticated algorithms alone. Aggregated models tend to obscure the idiographic, within-person dynamics that characterise authentic learning processes (Saqr & López-Pernas, 2025). By integrating theoretical perspectives, we can develop analytics that respect these dynamics, linking cognitive, motivational and affective dimensions to measurable patterns of behaviour. In this way, theory acts as a bridge between educational psychology and computational modelling, ensuring that personalisation strategies are transparent, fair and pedagogically interpretable.

Based on these insights, the LAK26 workshop aims to establish a community of researchers and practitioners dedicated to anchoring personalisation in learning theory. Participants will co-design models connecting constructs such as motivation, self-regulation, and cognitive load to learning analytics indicators; explore within-person and idiographic approaches; and articulate ethical workflows for tailoring feedback. By emphasising explanation over correlation, this initiative contributes to LAK's broader effort to strengthen the conceptual and methodological foundations of the field, ensuring that personalisation in learning analytics remains scientifically rigorous and educationally meaningful.

References
Bourdieu, P., & Wacquant, L. J. D. (1992). An Invitation to Reflexive Sociology. Polity Press.
Cronbach, L. J. (1957). The two disciplines of scientific psychology. American Psychologist, 12(11), 671–684.
Cronbach, L. J., & Snow, R. E. (1977). Aptitudes and Instructional Methods: A Handbook for Research on Interactions. Irvington.
Gašević, D., Dawson, S., & Rogers, T. (2016). Learning analytics should not promote one-size-fits-all. The Internet and Higher Education, 28, 68–84.
Kalyuga, S. (2007). Expertise reversal effect and its implications for learner-tailored instruction. Educational Psychology Review, 19, 509–539.
Panadero, E. (2017). A review of self-regulated learning: Six models and four directions for research. Frontiers in Psychology, 8, 422.
Saqr, M., & López-Pernas, S. (2025). Changes in online engagement at the within-person level. The Internet and Higher Education.
Wise, A. F., & Shaffer, D. W. (2015). Why theory matters more than ever in the age of big data. Journal of Learning Analytics, 2(2), 5–13.