Learning analytics is a new area in the assessment domain of learning and is based on interrogating large data sets of student information with a view to improving student retention in courses and student academic performance. Learning analytics is becoming very accurate in predicting student success. However, there are issues about collecting and analysing large data sets of student information that need to be considered in an educational context before learning analytics systems are introduced into higher education.
A recent thought provoking and very useful article Ethics, Big Data, and Analytics: A Model for Application addresses questions of responsibilities associated with the data mining of student information. ‘Systems now exist that link students’ demographic data with data on past and current performance, engagement with online courses materials or in-class participation levels to predict, with startling accuracy, specific outcomes such as final grades within a course’ (p. 1). However, using learning analytics based on big data means that issues such as ‘what it means to “know” with predictive analytics’ (p. 1) become critical to ensure educational integrity. For example, ‘What is the obligation to act on information?’ In an educational context, relations between students, lecturers and institution administrators have a learning focus that values academic success and so acting on information using learning analytics becomes sensitive for all parties involved in learning.
The use of learning analytics enables the participants to gain new insights into learning behaviour and performance, that is, ‘which students are performing well and which students need help’ (p. 2). A case study of a university that has implemented learning analytics is especially useful. The university has collected student data on ‘student demographics and academic history, engagement with online resources, and current performance in a given course’ (p. 2). This information has enabled the university to predict the students at risk and to provide additional support and assistance. The results of the program have been that the student retention rate has increased (from 69% to 93%) and students graduated at a higher rate (p. 3).
The benefits of learning analytics are its predictive value, its capacity to provide feedback to students and also to provide targeted support and additional resources where necessary that have maximum effect. Who has the responsibility for student performance in this situation? Is it the student, the staff member or the institution or do each have a partial responsibility and what access to information should each have? These questions become issues for the university, the staff and the student that can cut across university policies and privacy laws.
Ethics, Big Data, and Analytics: A Model for Application provides a model for examining these issues based on a university case study. The model that is proposed focusses on how to use learning analytics to collect and analyse student ‘information about the likelihood of academic success in a meaningful and effective manner’ (p. 3). Improving student retention and success is an important goal for academic institutions, and so training and development in the use of learning analytics tools is necessary for staff, so that they can provide timely feedback on students’ progress.
In an environment where learning analytics is used to improve student learning, there are a number of responsibilities that the institution, staff and students must accept. Ethics, Big Data, and Analytics: A Model for Application outlines those responsibilities and the benefits of learning analytics very well indeed.
Willis, J. E, Campbell, J.P., & Pistilli, M.D. (2013). Ethics, big data, and analytics: a model for application. Educause Review, Mar/Apr. Retrieved from http://www.educause.edu/ero/article/ethics-big-data-and-analytics-model-application
Image: Flickr /Photo Extremist
Gerry White is Principal Research Fellow: Teaching & Learning using Digital Technologies, Australian Council for Educational Research
This article was first published on the Australian Council for Educational Research's Digital Education Research Network 2 (DERN) and is reproduced here in full, courtesy of DERN.