When Are Learning Analytics Ready and What Are They Ready For

Abstract

Learning Analytics as a field of inquiry and community is distinct in the way that it brings together in shared pursuit, the research and practice of a particular kind of educational technology. At times this relationship approaches symbiosis: the annual LAK conference offers opportunities to learn both about the latest theoretical, methodological, and technological innovations as well as challenges and effective strategies for using such innovations to support learning in real world contexts. At other times, we feel pulled in multiple directions by the different priorities of each endeavor. Research is first and foremost concerned with advancing the state of the field by building knowledge, theories, techniques and tools with generalizable implications. Practice is primarily focused on action and implementation to have a positive impact on real world learning contexts. While these aims are not unrelated, they often offer quite different answers to the question of when a learning analytics application is ready to be used in an authentic educational setting with actual learners (and real consequences). At the extreme pole of a research perspective, there is always the temptation to try one more way to optimize an analytic (have we considered all possible features, tried all appropriate algorithms, tweaked all available hyper-parameters, explored all possible visualizations etc.). But from the perspective of practice with pressing problems to address, a tool that is available and sufficiently optimised is better than an unavailable perfect one. The ultimate goal is increased impact on learning not simply improved model accuracy. Of course, as a scholarly pursuit, the field of learning analytics does not seek to only develop and implement innovative data-based technologies, but also develop a knowledge base around them. Thus the key question is how can we make a difference in the world while also engaging in a rigorous knowledge producing process?