I was at Britto’s, the famous bar and restaurant in Goa, with a bunch of friends waiting to savor some fish curry, when one comrade at the table sparked out a debate about the perils of generalization. The group was talking about people and stereotypes in general when this close friend of mine argued “…but how can you generalize? Not all people from so and so country or so and so community act in the exact same manner and your generalized view about them is unfair!” “I know…they don’t!” I argued back “and yet there are some general trends about them…” Our argument could’ve easily taken a serious turn had we not been on vacation and at that point in time, starving! So, when the food arrived, we decided to dive straight in (to the food) and thereafter continued to joke about our argument throughout our trip, by prefixing every other statement with “generally speaking…!”
The fun argument with my friends stayed with me even after I returned from Goa more so because this very question of “how can we generalize if what we are aiming at is to individualize?” had been on my mind from the time I began reading Big Data – A Revolution That Will Transform How We Live, Work, And Think – a book by Viktor Mayer – Schönberger and Kenneth Cukier. In the very first chapter of his book, the authors inform their readers about a key shift (amongst others) in thinking when it comes to big data and that is about losing out on the exactitude. The author states that with small amounts of data, the results are more accurate but the bigger the data gets, the more the inaccuracies.
I am reading about big data because I am curious about learning analytics, which in turn relates to adaptive learning. And my interest in the topic is from the perspective of learning design – how does learning design draw from learning analytics to create an individualized path for the learner? Incidentally, BJET, the British Journal of Educational Technology is calling for white papers for its special issue on integrating learning design, teacher-led inquiry, and learning analytics, which will be published in January 2015. I am oh so looking forward to finding many answers in the literature that will emerge from this initiative.
So, coming back to my question— can learning analytics lead us to the preferences of the individual student? Or does it only create categories and slots each student into a category? In effect, are we still looking at generalizing? Although, this type of generalization is still better than “one size fits all,” but does this generalization lead to the individualization that is possible via a one-on-one between the teacher and student?
A couple of years ago, I worked on a project that involved developing an online degree program for students with Autism Syndrome Disease (ASD). And the USP of this program, as promoted by the college, was an “individualized learning path” for each student, while maintaining academic integrity. The key aspects of the program that contributed to individualization included the following:
Content delivery in multiple formats to suit a variety of learning styles
Flexibility in assessment and participation modalities
One-on-one support from qualified behavior analysts as mentors who act as liaisons between the faculty and the students
Each mentor would be assigned 3-5 students and then the mentor would work with the faculty to customize and individual all aspects of the learning path (including administrative and use of technology) for the students.
Can learning analytics help achieve this?
Although the broader benefits of learning analytics are directed toward bringing in efficiency to the teaching/learning process and to enhance the quality of education ensuring learner success, personalization of the learning process, instructional design, and content is achievable, provided the learning materials are designed to reflect the knowledge architecture of a domain.