The power of small data in education
As the tech industry's capacity for gathering and parsing big data has grown rapidly in recent years, its collective imagination is working overtime creating opportunities to put that data to work. Netflix uses deep learning to find out whether I'd prefer to watch “Bering Sea Gold” or “Cake Boss” while I fall asleep. Google looks for cues in my searches on, say, snow conditions to get ski boot ads at the top of my screen. And Amazon, according to its recent patents, is even looking into "anticipatory shipping" to have its army of drones ready to ship products before I've even clicked purchase.
A lot of effort has gone into finding avenues for big data recently in education, as well. Some college administrators now claim their big data algorithms predict which applicants will be “high yield” students before they've even been accepted. On the other side of the grade book, two education tech companies claim predictive analytics can show how successful a teacher will be before he or she even enters the classroom.
Perhaps most notably, we've seen how the “No Child Left Behind” act and similar initiatives that place all their emphasis on metrics are leading to an over-centralized, no-datum-left-behind approach to education. But do we really want the state boards of education playing “Moneyball” with our schools? Kids aren't data points, or trends, or objectives; they're individuals with unique strengths, weaknesses and interests, and they deserve individual support.
The conversation we need to be having is around the role that technology can play in empowering teachers to bring more personalized instruction to each and every student in the classroom. That’s where the battle for the hearts and minds of our kids is being fought, not in the board rooms or legislative chambers. For all its promise, big data doesn’t affect what happens in the classroom day in and day out — it’s really all about what we call the small data. Supercomputers crunching end-of-year assessment averages across the state have little to offer on the topic of whether or not a given student understands ratios. But a teacher using technology to monitor student understanding in real time can make the difference.
Let's back up about a century: Our education system was born out of a series of arbitrary decisions that led to a factory approach originating in the industrial revolution and characterized by evaluating learning at the end of an instructional unit, or what educators refer to as “summative assessment.” The most notable problem to this approach has been the lack of immediate feedback and remediation on concepts that a student may be struggling to learn. An alternative based on continual monitoring and remediation, called “mastery learning,” was introduced in the early 1900s, but the time required to implement this kind of approach always precluded its widespread adoption.
In the years since, the mastery learning philosophy, also referred to as formative learning, has been revived at various times by educators and philosophers like James Block and Sal Khan. As Dylan Wiliam noted in a research brief for the National Council of Teachers of Mathematics, “The available research evidence suggests that formative assessment produces greater increases in students’ achievement than class-size reduction or increases in teachers’ content knowledge.”
Mastery learning has not been lurking on the fringe because its efficacy is debated, but because we just haven’t been able to pull it off. The key to mastery learning is being able to observe and measure concepts of learning and student progress while providing interventions in real time for concepts not mastered. This is where technology and “small data” come in — they're the missing links to enabling teachers, students and parents to observe and collaborate on vital learning outcomes.
Suppose you're a 6th-grade teacher and you've got 30 students in your class. Half of those students are missing some crucial skills in geometry. A handful of others participated in an accelerated math program and are ready for algebra. And then there are a few others still who have learning disabilities and will need plenty of one-on-one time to progress. Ideally, you'd have time to sit down with each student and craft a teaching plan specific to his or her needs and goals. Unfortunately, you're handcuffed by prescribed lesson plans, fixed test dates, and an inflexible grade book that doesn't truly distinguish between who understands math standard 6.RPA.1 and who doesn't.
This is why, in an interview with the Harvard Business Review, Sal Khan said, “Historically, that was very hard to do in a scalable way, where, hey, how do you personalize education for 30 kids and do it without breaking the bank? And that’s what’s exciting about technology, is that the technology can do a lot of the information delivery at a student’s time, at a student’s pace, can give practice problems, give feedback, arm teachers with data.”
Finally, cloud and mobile technology have matured to the point that mastery-based assessment is a viable option. What were once curriculum assessment systems are now becoming Mastery Learning Systems that are giving teachers the ability to analyze the performance data of their students and then, most importantly, do something about it in real time.
These technologies expedite the work of personalizing education on a massive scale, so that, instead of focusing on managing a class of 30 students, teachers easily identify individual students who are having trouble with certain concepts in actionable detail. Teachers can target them for more individualized attention or extra practice at home with parents, or more interaction with the teachers themselves. By the time an assessment arrives, teachers and students have a more precise sense of how ready a student is to move on to the next unit.
Contrary to big data, which puts the information of many into the hands of a few, “small data” puts the information of the few into the hands of the few so they can take advantage of it. We're on the verge of mastery learning gaining widespread adoption thanks to a growing appetite for real change and a mature technology infrastructure to make it possible. Now it's time to shutter up the old factory and embrace the one-on-one instruction that is finally within reach.
Image via Wikimedia.