It’s being said that the future of education is machine centered and algorithmic, and the greatest critique of this vision centers on a lack of transparency. ( Waters and Williamson). If we want to understand the impact of Ed Tech and Big Data, as well as to shape our own future, we should start with clarity; a clear eyed view of who we are, who we want to be and the pedagogical processes to get there. That is, let’s specify the ontology (who is an educated person) and the teleology (developmental pedagogical processes) of principled and well-structured Ed Tech information systems designed to serve the educational needs of networked people in dialogic relations.
An ontology defines a common vocabulary for researchers who need to share information in a domain. It includes machine-interpretable definitions of basic concepts in the domain and relations among them. . . . There is no one correct way to model a domain— there are always viable alternatives. The best solution almost always depends on the application that you have in mind. (Noy & McGuiness, )
This ontology should reflect evidence-based competencies, not just the parroting of knowledge.
When higher level skill sets are the real objects of measurement, it is necessary to evaluate assessment activities not by their surface similarities with learning domains but by their deep structural correspondences with intended learning outcomes; . . . To ensure that assessment activities yield useful data for making inferences about student learning beyond simple knowledge claims, principled assessment design must guide the development and structure of the assessment. Principled assessment design can be viewed as a plan, comprising a visual or textual scheme, to guide the purpose, expression, development, internal structure, and defensibility of an assessment. (Shute et. al. 2014)
If we don’t achieve specificity, algorithm designers will continue to do it for us in opaque and thoughtless ways. I believe that transparency problems can exist not only because of interference with corporate interest as Dr Williamson implies. but also because of a lack of clarity in principled system design. Specifying the underlying ontology and teleology of Ed Tech and Big Data Systems will go a long way to improving this situation.
For further clarification, ontology in machine learning can be seen as different from ontology in philosophy, but when we look at these applications as educational processes, we need to look well beyond the code. A common refrain in Ed tech is that the field is populated by programmers with little understanding of the history and concepts of education. This is to say that programmers think of ontologies and applications as limited to the current program code, but educational applications should reflect networked people in dialogue. In defining a common vocabulary, an ontology’s domain should support students, student development, and the the educational process. This aspect forms the core ontological commitments that allow a model of the domain to be created in a way that is meaningful across the domain for teachers technologists and students. This ontology is also important for interpreting the analysis and applying the data analysis to the process of educational and personal student development. Without an interpretation that also reflects ontological commitments it can’t fully communicated and implemented in the kind of educational practices we expect today.
(T)he goal of data collection and analysis is to provide insight and inform decisions. Accordingly, there is a long chain of reasoning that needs to be considered.” We recognize that data is a representation of the world and like all representations, it is an imperfect system which will not perfectly capture the detail of the world. We also believe that all of the activity coming after that (analysis, interpretation, etc.) is a human endeavor, involving all the benefits and challenges that implies. (Kristen DiCerbo, Pearson)
DiCerbo provides a chain of reasoning that lacks an ontology and teleology. “Big Date” in this view is not based on a principled assessment design. What will result is much more than an imperfect reflection of the world, but an opaque data system. What we need is more than people with knowledge on the inside. What we need is principled assessment design backed up by principled system design. More than just trust, more than just efficiency, we need systems that are worthy of guiding educational teleology.