Background Information
The problem
What are the grand challenges of Learning Analytics (LA)? Where is our theoretical contribution and what specifically are we adding to the field of education? Although some attempts have been made to highlight how LA might address large scale challenges (Buckingham Shum, 2023) and indeed, a list of grand challenges for the field has been put forward (Baker, 2019), we are yet to coalesce around a community-defined set of research priorities. Against this backdrop, the aims of large LA research groups are not always aligned, and there have even been recent bandwagon effects where a ‘hot topic’ emerges and distracts attention from areas with potential for benefitting the field. Most crucially, new entrants to the field and early career researchers sometimes find it difficult to understand the rich background landscape of the field, and why certain problems have been identified as important. Without a clear unifying set of challenges, it is likely that LA will make only incremental contributions, if any. We are at risk of becoming feudalistic, with various teams staying within their safe, identified subfields.
While education itself is often touted as a field that will help us to create a more equal and just society, LA is sometimes accused of supporting agendas that will track people, violate their privacy, and manipulate them towards acts that they might not have undertaken on their own. How can we work towards ensuring that the field is solving big issues that help to ensure the next generation of people are more mindful and accepting of each other and the differences between us, respond less to the abundance of false information, and are able to adjust in ways that are well reasoned rather than simply reactive to societal shifts and new challenges?
The solution
A number of research disciplines have coalesced around an established list of grand challenges, resulting from periodic workshops bringing together key members of that field. Perhaps the most famous effort in this space was instigated by Hilbert, who in 1900 proposed a set of 23 unsolved problems in mathematics that drove research throughout the 20th century (indeed they still do as only 15 of them have been solved to date). This programmatic approach to defining a field has inspired other research domains. For example, the field of Information Retrieval has held three SWIRL workshops (in 2004, 2012, and 2018) where leading figures in the field were invited to attend and define challenges and opportunities for the field. This type of prioritization can both guide future research for new entrants, and support proposals for funding and tenders.
Similar attempts have been made in the educational sciences, but the results have yet to achieve broad impact upon the directions of LA as a field. For example, in 2011 a STELLAR workshop organized as a part of the Alpine Rendezvous series (Mwanza-Simwami, et al. 2011) proposed six grand challenges for Technology-Enhanced Learning (TEL) in Europe but to date the white paper that emerged has received zero citations which leads us to believe that despite deepening the scholarly discussions in this area within the interested community, its broadscale impact upon the field has been limited. So far, none of the proposed lines of work has been pursued to the point of completion or resolution, despite the paper listing likely timeframes and measurable indicators of success. While we believe such workshops that identify grand challenges hold potential for disciplinary scholarship through collaborative development of priorities, the question of how specific the challenges of LA might be, and how they differ from those of TEL remains unanswered. Grand challenges that pertain to artificial intelligence in education (AIED) have also been pulled from other branches of computer science (Kay, 2012) in a process that, while useful for defining relevant challenges and sparking research, has not necessarily been organic to the AIED community itself. Furthermore, their specificity to AIED marks those challenges as potentially too restrictive for our field. Baker’s (2019) grand challenges have gained considerable attention (they have been cited 129 times), but are particularly aimed at educational data mining. While they might represent a starting point for discussion, we do not consider them representative of the work occurring in the broader LA community, leaving an opportunity that this workshop will address.
It is time for the learning analytics community to “expand our horizons” by collaboratively defining the problems and opportunities of the field. This workshop will attempt to bring together a range of stakeholders with different voices and backgrounds to define a set of community-accepted and supported grand challenges that can drive the next 10 years of LA research and development.