@incollection{knight_theory_2017, address = {Alberta, Canada}, edition = {1}, title = {Theory and {Learning} {Analytics}}, isbn = {978-0-9952408-0-3}, url = {http://solaresearch.org/hla-17/hla17-chapter1}, abstract = {The challenge of understanding how theory and analytics relate is to move “from clicks to constructs” in a principled way. Learning analytics are a specific incarnation of the bigger shift to an algorithmically pervaded society, and their wider impact on education needs careful consideration. In this chapter, we argue that by design — or else by accident — the use of a learning analytics tool is always aligned with assessment regimes, which are in turn grounded in epistemological assumptions and pedagogical practices. Fundamentally then, we argue that deploying a given learning analytics tool expresses a commitment to a particular educational worldview, designed to nurture particular kinds of learners. We outline some key provocations in the development of learning analytic techniques, key questions to draw out the purpose and assumptions built into learning analytics. We suggest that using “claims analysis” — analysis of the implicit or explicit stances taken in the design and deploying of technologies — is a productive human-centred method to address these key questions, and we offer some examples of the method applied to those provocations.}, booktitle = {The {Handbook} of {Learning} {Analytics}}, publisher = {Society for Learning Analytics Research (SoLAR)}, author = {Knight, Simon and Buckingham Shum, Simon}, editor = {Lang, Charles and Siemens, George and Wise, Alyssa Friend and Gaševic, Dragan}, year = {2017}, pages = {17--22} }