Learning analytics looks are proxies for learning, and it can be tempting to mistake correlations for causation.
Predicting the future is an enticing idea to academic leaders. Programmers have their “Design Patterns”, which are methods to solve problems in the code, before such problems become evident. However, we have seen some unfortunate stories connected with analytics and Big Data in education. Mount St. Mary’s University in Maryland “drown the bunnies” is a good example. Even students are calling out to have permission to fail. I think we should welcome failure and use it as a tool for teaching.
Quantitative research in education dealing with student success is not objective. Studies funded by textbook publishers find that students who read the textbook often get better grades. LMS companies report that students who check their grades often are more successful. May there be a causation versus correlation issue here?
Recently Blackboard Inc posted an interesting article about analytics. There was a paragraph on “downsides”. I appreciate this thread of concerns being voiced in published reports.
“Learning analytics looks are proxies for learning, and it can be tempting to mistake correlations for causation. Learning analytics requires close cooperation between campus departments that traditionally have worked independently (e.g., IT, academic affairs, student affairs, and faculty). Data required for learning analytics can be distributed across campus and difficult to integrate, particularly if technology vendors format data in proprietary ways. Available data may not be suitable for analytics models. Using student data for analytics raises ethical issues surrounding data privacy and institutional obligations to act on analytics findings, including by providing resources to assist those learners. Analytics algorithms may include biases and may mislead the very students they are intended to help, perhaps prioritizing efficiency toward a credential over a learner’s passions. Misapprehensions about analytics among university administrators can result in unrealistic expectations for results, and some faculty resist analytics, arguing that it focuses on behavior rather than on learning.”