Background

This workshop is founded on the premise that the quality of learning analytics, both research and practice, rests on the strength of its connection to theory (Gašević, Dawson, & Siemens, 2015). Through this workshop we hope to build an ongoing community of scholars interested in both using educational (and other) theory in learning analytics research and practice, and contributing to further development of theory through their work.

Theory provides a common language through which to communicate about research, it gives a frame of reference to understand the type of knowledge being generated, and what may be legitimately claimed (Reimann, 2016). In a typical research cycle, we suppose that theory influences the questions we ask, design of data collection, analysis approach and method, and interpretation and reporting of results (Wise & Shaffer, 2015). In this way we are arguing for a move away from the primacy of method in learning analytics, that is, away from pragmatism to theory-driven paradigms for research where theory underpins method and the two cannot be separated (Bartimote, Pardo, Reimann, 2019). This adds the possibility for explanation – for an observed pattern, for a prediction, for why an intervention or pedagogical strategy works – in research, and in practice.

Theory allows for informed practice by a range of actors that support learning in educational settings, such as teachers, student support officers, advisors, and academic managers. If the objective of learning analytics is actionable information, then theory-driven analytics enables choices and decisions that are situated in defensible frameworks (Bartimote, Pardo, Reimann, 2019). And it means we have a starting point for explanation when things do or don’t work, and a basis for adaption of tactics and strategies shown to be effective in one context, in other contexts. For analysts, data scientists, and software developers, theory may guide what activities to capture, the development of indicators and measures, the display of information, and the form of personalised messages and automated nudges. We need to focus on providing information about constructs that matter, and learning (and other) theories substantiated by empirical research can serve as useful starting points.

The LAK community is increasingly drawing on ideas from the learning sciences, educational psychology, sociology, and social psychology. This is demonstrated by the inclusion of an educational theory session on the LAK 2019 program, but more generally in recently published learning analytics work referring to theories such as social cognitive theory and self-efficacy beliefs, various self-regulated learning models, measurement theory, social-constructivism, human-computer interaction (HCI) and activity theory, Kolb’s experiential learning cycle, etc. We consider the time is ripe for a call across the community to gather to consider more explicitly the role of theory in learning analytics.

Moving on from the LAK 2019 Educational Theory session, it is necessary to address some of the issues raised regarding definitions of concepts, design, model validation and interpretation of findings. To do this, multidisciplinary groups of researchers working in the area need to come together to support this work and begin to create some level of understanding in the field. This is the work proposed for the LAK 2020 theory workshop.

REFERENCES

Bartimote, K., Pardo, A., & Reimann, P. (2019). The perspective realism brings to learning analytics in the classroom. In J. M. Lodge, J. Cooney Horvath, & L. Corrin (Eds.), Learning Analytics in the Classroom: Translating Learning Analytics Research for Teachers. London: Routledge.

Gašević, D., Dawson, S., & Siemens, G. (2015). Let’s not forget: Learning analytics are about learning. TechTrends, 59(1), 64-71.

Reimann, P. (2016). Connecting learning analytics with learning research: The role of design-based research. Learning: Research and Practice, 2(2), 130-142.

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.