Smart Tools, and Not Just for Smart Kids
By Carole Beal
Leo Tolstoy famously began Anna Karenina with the observation that "All happy families are alike; each unhappy family is unhappy in its own way." While it's an overstatement, this also applies to K-12 education where "all good students are alike; each failing student fails in her or his own way."
We more commonly call this phenomenon the "rich get richer " effect. Seemingly exciting new educational tools and products characteristically work well for students who are already superior but do little or nothing for students whose work is below grade level or who are failing outright.
What if we could quickly and accurately diagnose how each failing student fails and provide them with specific tools to address their specific pattern of failure?
Inspired teachers know how to do this: that is why they excel. But teachers are under multiple constraints and heavy pressure to deal with large classes in the 180 days per year they have available for teaching. Even great teachers need all the help that they can get, and many of them are asking for technology-based tools to help their most needy students.
Information technology can provide this support, with appropriately designed teaching materials -- the kind we're now building. What is critical in these systems is not so much that they teach certain educational materials as the way they engage with individual students to see where their learning problems are.
One example we've developed is "Wayang Outpost," a multimedia learning system meant to teach geometry to high school students. It seems to work, bringing a 20% improvement in post-test scores. But it improves scores only on the skills taught in the system, not on other material, so we know that the benefits are not due to a transient novelty effect. The weakest students post the biggest improvement, which is a reversal of the usual “rich get richer” effect that is so often found in the education community.
Wayang doesn't just give vivid multimedia explanations of geometry. It lets a failing student try a problem, fail, get immediate feedback, and learn from failure. Only the computer knows about the failure. It helps discouraged students while shielding them from embarrassment in front of classmates or the teacher. Our research indicates that many of these “failing” students are eager for such opportunities.
The key is that Wayang studies and learns from its users. With data gathered from the student, the software creates a model of where each student is -- spatial ability, math proficiency, and prior knowledge. The system surveys their motivation from a variety of angles (their learning style, test anxiety, self concept in math, and even their mood on that particular day) and produces a detailed picture of each learner, classifying them into various learning categories.
It tailors an educational program for each student using information about how earlier students with similar profiles had done in the system. It can recognize bad "trajectories" such as the student forming an anti-successful strategy to get by or coast through, and it intervenes before the student is set on the track. For example, Wayang can diagnose that the student is guessing, and remind the student of his or her goal to go to college, making the link between today’s class work and where the student wants to be in a year.
Information about the student comes not just from what he or she types. New technologies allow us to monitor the student and do so with a “light footprint”. Simple, inexpensive web cameras can track eye gaze to see when attention lags. We’re starting to experiment with lightweight wireless EEG headsets to directly measure cognitive effort and distractibility.
It is not just students who can profit from these technological individualized tutoring tools. We believe the tools will also help teachers learn to be more effective in the same way. The systems provide teachers with real-time assessment about each student’s strengths and weaknesses – feedback they can use to adapt the next day’s lesson plan – instead of waiting for the results of annual achievement tests.
Unfortunately, technology has often been oversold in education so I don't want to exaggerate where we are. We still have a long way to go. But I believe we are on the right track in focusing on technology that listens to and diagnoses students. Our goal is to provide each student with individualized tutoring -- emphasis on individualized. Information technology can do this.
Carole Beale is a project leader and educational psychologist specializing in K-12 education at the USC Viterbi School’s Information Sciences Institute. 
