As a person with a vested interest in a company that relies primarily on face-to-face, human-to-human teaching methods, I certainly hope robots never take over education. However, I do hope that adaptive learning technology continues to make rapid gains in its ability to help students learn more efficiently.
By “adaptive learning technology,” I simply mean any sort of education technology that changes its instruction based on inputs from the student. Clicking through a PowerPoint presentation or watching a video are not examples of adaptive learning technology. Instead, imagine a system that gives a student a diagnostic test, analyzes that test and then delivers a custom-tailored lesson that is highly targeted in the areas the student showed the greatest weakness. Repeat that cycle several times, and it’s easy to see how technology that “learns the student” could yield major gains.
To be clear, such a system is likely most effective when it serves as a tool to assist a real-life, human teacher. As a teacher, I can pick up on subtle cues from students that even the most advanced machines are currently unable to do. I can tell from a facial expression that things aren’t quite clicking yet or from a student’s body position that my lesson simply isn’t as interesting as it should be. There are some things I’m not very good at, though, and that is where adaptive learning technology could improve student outcomes.
For starters, humans aren’t great at analyzing large swaths of data. Computers have already been better at this than humans for decades. If you give me a complete set of results from a student’s practice test along with corresponding categories for each question, I cannot derive as much meaningful information from it as a well-designed software system could. This is an important function in the tutoring process, and having technology that can automatically process a student’s performance, identify areas of need and prescribe a plan to improve in those areas will mean a major improvement in the effectiveness of any educational program.
We’re currently in the process of designing such a system at Test Geek. It is challenging, but it promises to become a major part of the way we work with students. When our students take a practice test, they won’t simply get a score report that tells them how they did on each section and what their total score is. They will get all of that, but they will also see how they did in dozens of micro-categories within the test. A student knowing that she scored a 24 on the ACT Reading Test is only moderately helpful. Knowing that she did very well on questions that call for basic information finding but poorly on questions that require inferences is very helpful, however. If she then has a set of problems assigned that will specifically target inference problems, there is a high likelihood that she will improve.
That’s our goal with this system. We’re already working hard to have the best curriculum and the best instructors, but I believe it is this technology that will allow us to consistently deliver the biggest score improvements possible.