The Personalized Learning Model and The Game of School
Brandt Redd, the Senior Technology Officer in Education for the Bill & Melinda Gates Foundation, provided a new perspective on how technologies enable positive changes to simultaneously make school more fun and increase learning achievement. The current rules for school are designed around scarcity of teacher time and material resources. Learning maps, abundant assessment and feedback loops are three remedies that can free teacher time for more personalized learning. And these are what we can leverage the power of technology in education. His insights are shared under creative commons license as CC-BY.
The first two parts of this series discussed the Tyranny of the Bell Curve and a strategy for Tackling Bloom’s Two Sigma Problem. In this third and last part I describe the Personalized Learning Model that many of us are using to guide investments in education technology.
The diagram to the right is similar to those used by other education technology organizations so it’s not unique to the Gates Foundation. The key components in most any Adaptive Learning System or Instructional Improvement System are Student Data, Educational Content and Assessments. We use precise definitions of these:
- Learning Objectives are specific competencies to be learned in a particular subject domain. Most courses, both online and legacy media, start with a set of learning objectives. However, if data, content and assessment systems are to interoperate, a common set of objectives must be shared among them.
- Student Data is a collection of evidence of what competencies or skills a student has achieved. On a scale of weak to strong evidence, it includes presence information (the student attended a class), activity information (the student viewed a particular video or performed a lab) and assessment results.
- Learning Content includes reading materials, textbooks, interactive activities, lesson plans, exercises and any other content that’s intended to teach about a subject.
- Assessments are student activities that are instrumented in such a way that we can measure competence in knowledge or skills. You can think of multiple choice and true/false as activities that are deliberately simplified to make them easier to instrument. However,assessment technology is advancing in ways that make it possible to instrument more realistic activities.
The Feedback Loop describes the process of learning, from determining what a student doesn’t know, to teaching the subject, to assessing competency. For the feedback loop to work effectively, it must cycle frequently supplying rich and accurate feedback to students and educators.
Most of our education technology investments involve some combination of improving the state of practice in these areas and improving interoperability among systems. Future posts in this blog will profile some of the most important initiatives we and others are working on.
I come from a family of engineers. Naturally, we cannot talk about feedback loops without getting into control theory. Conveniently, engineering control theory can be adapted to education reasonably well as soon as you get over some vocabulary hurdles.
|An open-loop control system.|
Here is a basic control system. One example might be a cruise control for your car. In that case, the reference is the speed you want to go, e.g. 65 miles per hour. The controller translates that reference to an input value – the throttle position . Then the system (the engine, transmission, drive train, tires) produces an output – the actual speed of the car.
In control theory, this is an “open-loop” control system meaning that it has no feedback. The controller must have a very good mathematical model of the system and the system itself must be very precise to get a predictable output. Open loop systems are used when there’s a large acceptable margin of error. A cooling fan, for example.
|A closed-loop control system.|
A car with an open-loop cruise control would slow down as it climbed hills and speed up when it descended. For more precise control, real cruise controls use a feedback loop. A sensor detects the output value (the speed of the car) and returns it to the controller. The controller takes the difference between the feedback and the reference values and adjusts the input (throttle) accordingly.
Mathematically, taking the difference is subtraction so we call this “negative feedback.” In contrast, “positive feedback” sends systems wildly out of control. A familiar example is pointing a microphone at a speaker. The terrific squeal that you hear is a “positive feedback loop.”
So, when talking about control systems, “closed loop” is better than “open loop” and “negative feedback” is good while “positive feedback” is bad. The precise engineering terminology is the opposite of what we might use in casual conversation. Let’s now apply this to education.
|A personalized learning system.|
With a few word substitutions we get a personalized learning system. For a moment, let’s ignore the feedback loop. In an open-loop learning system a set of educational standards (such as the various state or common core standards) are translated into instruction in the form of textbooks, lectures and exercises. These activities are delivered to the student resulting in skills. Conventional schooling is much like this. In order to handle a class of a dozen or more students, all students perform the same activities. But, remember that open-loop systems require a very precise system to get predictable results. Students come to us with different personalities, talents, preferences and backgrounds. The result is a wide margin of error as illustrated by the spectrum of grades assigned at the end of of a course.
Good teachers do better than this. They find ways to adapt their teaching to the needs of individual students. Good students, being intelligent, can also use feedback to adapt their learning to match the instructional methods being used. Both kinds of adaptation require feedback from assessment. And to do this well, the feedback should be compared with the standards for what is intended to be taught.
For feedback to be really effective, it must be frequent, fast and rich.Feedback must come often so that course corrections are made frequently. It must be fast, ideally immediate; otherwise it’s too late to affect the learning process. And it should be rich. “Incorrect” is not nearly so meaningful as, “You misplaced the negative sign in step 3.”
Frequent, fast and rich feedback depends on frequent, fast and rich assessments. In conventional schooling, assessment is expensive. It costs a lot of teacher time to compose, administer and score assessments. This is an excellent opportunity to apply technology. Computer supports can minimize the effort required to compose, administer and score assessments. And computers can tabulate the results into effective teacher and student dashboards. The result is a more personalized and effective learning experience. Or, as an engineer might say, “more predictable output with a smaller margin of error.”