Learning analytics treats students as numbers, invades privacy, and doesn't elevate education.
When I was about six months old, my well-meaning and sleep-deprived parents bought me an indispensable and comprehensive "education technology": the multivolume, all-inclusive reference source known as the Encyclopaedia Britannica. The salesman convinced them that this "august repository of serious information", organised alphabetically by topic and arranged over 26 volumes, had an impeccable pedigree, an excellent reputation and a history of trustworthiness. The pitch worked. The "robust intellectual content" of the Encyclopaedia Britannica was theirs and my cash- strapped parents spent the next two years paying off their purchase.
Today, education technologies offer the 21st century parent, similarly concerned about their children's educational future, an almost infinite number of learning opportunities. From the World Wide Web to Wikipedia, from Moodle to MOOCs, this century has embraced the interweb as the new "multivolume, all-inclusive reference source". The rich collections of data associated with these "learning opportunities" are starting to interest educational entrepreneurs who see rich pickings in the process of analysing data and packaging it in sets of marketable information. Much like Amazon is able to look through all my online purchases and suggest a new good read, or Facebook is able to suggest friends by determining the degrees of separation between me and the rest of the world, the computational process of discovering patterns in large educational data sets (data mining) is being touted by educational entrepreneurs as a solution for group work, career choices, retention rates and a higher quality, personalised experience.
Learning analytics is the current all-encompassing term used to describe this marriage between big data and educational records. Collected data about students is being used to create models that will predict student progress and performance; and allow parents, students, teachers and administrators the ability to act on that information.
Big data is hooking up with school reporting, and learning analytics is their new love child. Somewhere, probably in Silicon Valley, there's a company or two coding an app that will collect all available data about students, including academic records, ANA test scores, financial circumstances, geographic location, attendance, assignments, progress, discipline, demographics and health. This company will soon be spamming school's or university's home (pages), claiming to offer uncertain high school students guidance counselling by aggregated data mining.
Forgetting the experts
My guess is that the people coding this app do not have many educators on the design team. As Encyclopaedia Britannica was a sales-driven company, and not a scholarly one, data miners are looking at data as their source of profit and tend to neglect the teachers and educational research. The absence of educators persuades developers that they are thinking outside of the box, but it is more likely the absence of expertise or research will lead to an app that exacerbates current educational problems.
Big data is hooking up with school reporting, and learning analytics is their new love child.
My family's Britannica sat on its own shelf in the lounge for years, and was used for the occasional school project, resolving points of debate and pressing flowers. Learning analytics may not be as benign. While it purports to be unique and offer a personalised information and guidance as they make various curricular and career choices, in education this conversation has taken place before.
Student tracking assigned students to different classes and different futures. Tracking began with pre-school, screening, IQ tests and other achievement tests designed to measure so-called "ability", and thus set in place an educational trajectory for 12 years of schooling. Do we really want to repeat the mistakes of the 70s?
Learner analytics also comes at the expense of teacher engagement. Increasingly stats-bound schools purchasing an app as described above will allow teachers to monitor what is happening from their desks, instead of taking the opportunity to engage in student learning and questions, coaching, facilitating and getting to know their learners.
Education is more than the collection of patterns of data. Learning analytics may promise unlimited and useful information on students, but treating students as numbers invades privacy and doesn't elevate education. It could lead to a watered down, prescriptive education system with little human or humanising engagement.
Encyclopaedia Britannica's ubiquity was because the company was built by "a culture of salesmen, not scholars". With the likes of iPads, the Khan Academy and MOOCs, I suspect that many parents, teachers, schools and faculties will soon also find themselves under siege from similar Britannica-like salesmen, this time straight from Silicon Valley. Their new pitch is an appealing one. They say educators can choose to use the data generated by learning analytics to improve the quality of students' prospects, or remain disconnected, and disadvantage their students' prospects.
There's no doubt that we need to make a serious effort to understand the processes of learning in the 21 century world. Learning analytics, however, cannot begin to reflect the complexity of these learning processes. Education cannot be reduced to a set of data patterns. Putting one's faith in educational entrepreneurs' ability to create another "august repository of serious information" is very much like my parents' choice to buy a Britannica.
* Derek Moore is a teacher and learning designer at eLSI, Wits University