Proceedings of the 5th Global Conference on Education and Teaching
Year: 2023
DOI:
[PDF]
Machine Learning and the Iron Cage of Objective Modelling for Educational Outcomes
Dr. Callum Philbin
ABSTRACT:
Middle and High School teachers’ use of formative and summative assessments are integral to teaching and learning, but there is a general understanding that these processes are inherently subjective. Despite their application as not being understood as particularly problematic in contemporary Educational Theory, there is an attraction to relying on standardised testing for more objective and automated quantitative methods that are aided by machine learning. Relying on various studies, machine learning techniques can be viewed as assisting in measuring educational outcomes, while still requiring educational stakeholders to interpret the data and work to support student progress. The application of machine learning in predicting or measuring academic performance should be viewed in the context of encouraging ‘data-informed’ educators, as opposed to ‘data-driven’ educators, and highlight the continued importance of the educator in the educational process. This research applies DiMaggio and Powell’s conception of the Iron Cage (1983) to explore the proliferation of machine learning for measuring academic performance within the Global Education Reform Movement. The paper argues that machine learning has a place for measuring academic success in global education but is better understood as an art, as well as a science, and in turn, should be viewed as one method of many diverse quantitative and qualitative approaches to understanding educational outcomes.
keywords: machine learning, educational outcomes, iron cage, global education reform movement, accountability