By Steve Hill, Chief Technology Officer
This post was originally published on LinkedIn.
The world of educational data has felt stuck for a while. Today, it is largely two-dimensional — The SIS and LMS have dominated the landscape of educational technology and thereby limited the data, the means, and the depth of meaning we have available to reason about learning.
While there are integration standards which have facilitated making this bridge more robust, it is still a narrow bridge. There are important data elements which flow across it, like enrollments and grades. Each end of the bridge has seen analytics offerings emerge, mostly from the vantage point of each respective source system.
Each space has dominant players entrenched in their respective markets, and integration between these systems has muddled through 3 generations of building narrow bridges of data exchange. The stark reality is that more than a few institutions still employ batch mode uploads of spreadsheets to facilitate enrollments and grades being synchronized between the LMS and SIS.
Within the LMS environment, LTI allows for the launch of extension elements in the learning experience. However, capturing learning activity from these extensions has been very limited in promoting deeper reasoning so we have at best limited reasoning about learning efficacy.
Analytics as we currently know it in Higher Ed is all too often limited to localized reporting from the LMS and SIS respectively. Progressive institutions might also be doing CRM reporting. The ambitious have negotiated bespoke data extracts from the respective providers and created a data warehouse or begun a big data initiative.
But still much of what passes as analytics today is operationally oriented and enterprise focused. It is coarse-grained in its domains, concentrating on aggregates like grades and attendance. Less commonly employed but portending to the future, some providers offer institution-tailored “At Risk” predictors, and engagement metrics.
We Lack An Educational Data Architecture
In effect, what we have today is an educational data architecture that is driven by sources rather than needs. To be fair, we have seen reasonable advances in the enterprise perspective of educational data, but we have yet to see this progress graduate to learner-centric models of personalization & enablement.
There Is Hope
There are signs that the data is fighting to break free from these constraints, or more specifically that a good many people feel this needs to happen. Emergent technology companies have appeared which seek to disintermediate the standing markets value proposition. Pure-play predictive analytics offerings have emerged which seek to ameliorate the hurdles of data integration from the various islands, providing a new global and longitudinal perspective. Marketplaces for learning objects have established a small patch of land in the hopes of enabling community-driven and curated libraries of learning experiences. Players in adaptive learning seek to get underneath the learning data within the LMS in a way that shifts the import of each respective player — is the long-term value the LMS as “player”, or in the Learning Intelligence for real-time optimization of the educational experience?
At The Pointy End
At the edge, a precious few have sought to establish new models of describing learning activities and unifying the stream of data across learning objects. It is this edge which I find most interesting. At the edge lives the question, what is the right structure for technology to best enable data-driven improvements in learning, engagement, and intervention? Is it finally time to rethink the system landscape around efficacy, efficiency, and a renewed empiricism focused on learning? Do we need a holistic rethinking about how education data should work, one that enables an ecosystem of innovation at the edges of the substrate? Do we need to stop, and demand a real educational data architecture to evolve from the silos and the bridges?
The future in my opinion is twofold: (1) a learning componentry which enables a greater focus on the learning; (2) a learning-data-focused substrate, a data circuit if you will, which allows for a natural unification of data via integration through protocol. The combination of the componentry and the circuit will enhance our capacity for reasoning by pushing us to be fine-grained about the evidence of learning, embracing multiple dimensions (n-dimensional) of the academic experience. The unified data path will allow us to create contextual perspective-based analytics, therein enhancing decision making of the student, instructor, coach, and academic advisor.