Free Downloadable Resources
How do we learn?
In 2021, EDUCATE Ventures Research collaborated with the Action Lab at the USA’s leading university in innovation, Arizona State, to try to understand how its students were ‘learning to learn’ (LTL). Learning is a definition continually debated, but most scholars agree that learning is a process that involves change that follows experience (Schunk, 2012), and that for the most part, it is internal, and invisible (Lefrançois, 2019). These characteristics have direct consequences on the ways in which learning can be evaluated. This Byte-Size piece is based on research that examines the jump from a university caring about students’ learning journeys to utilising learning analytics to investigate how those journeys occur in practice, and specifically focusses on the process of creating and using an ontology as a tool to connect the theoretical aspects of LTL with the data sources typically available in education institutions.
To prepare young people for the 21st-century workforce, some universities aim to improve their graduates’ 'learning to learn' skills. Learning to learn (LTL) is a highly complex and contextualized skill set. LTL begins way before students arrive at university, and is affected by their countless daily interactions with other people and with their learning environments. This chapter explores a case study of one such university, Arizona State University, which crossed the line between simply caring about students' LTL journeys into using learning analytics to examine how they occur in practice. Crossing that line requires bridging at least three gaps: the gaps between (1) evaluating LTL and explicitly integrating it into the curriculum; (2) evaluating LTL and universities' assessment regimes, and (3) the data that is typically collected about students and the data that is required to evaluate LTL.
In this chapter, we focus on the third gap, which is related to data collection. We describe the process of creating and using an ontology, as a tool to connect the theoretical aspects of LTL and the data sources that are typically available in education institutions. We then share some results from an analysis informed by the LTL ontology. Specifically, our approach combines process mining with more static data analysis, to address both the temporal and the contextualized nature of LTL. We report on the hybrid models that were developed, and we reflect on how the data collection gap could be further addressed. Our results indicate that LTL can be treated as a dimension in its own right, with a moderate relationship to academic performance. LTL does not seem to develop at an even pace but demonstrates patterns across each term, irrespectively of the learning design and teaching mode.
About the Authors:
EDUCATE Ventures Research's AI & Data Science Team
Professor Rose Luckin's EDUCATE is an accelerator programme that helps EdTech startups and educators access the research and evidence they need to unlock the greatest educational solutions. All the work is underpinned by the EDUCATE Golden Triangle - bringing together educators, EdTech creators and researchers.