Scaling Up Personalized [sic] Education

Creating a personalized education program presents many challenges. It presents even more challenges if you want your delivery system to be a computer. And it presents even bigger problems if you intend to scale it up.

The dream for many education reformers is not just to come up with a new system, but to manufacture a new system that can scale up and reach a broad market. Personalizing education for a single student requires a great deal of work and it can be hugely beneficial for that student-- but it's not very profitable.

So who wants to do the assignment this time?

But scaling up is best accomplished by standardization, and standardization is the enemy of personalization.

Consider a burger chain, one that has decided to become the anti-McDonalds. This chain wants customers to know that when they order a burger, they can have it their way. The chain will gladly hold the pickle or the lettuce, because they have developed a production system that easily allows for that sort of personalization.

But that personalization falls within a very narrow range. You can hold the pickle or the lettuce, but you can not have a black bean patty in place of the hamburger. You cannot have thin-sliced sirloin covered with mushrooms or avocado slices. If your tastes do not fall within the range of burger possibilities, if you are an outlier, you will take your business elsewhere.

This makes sense for the burger chain. There may well be customers who want, say, a herring burger with sautéed kelp on top, but those customers are so few and far between that it's not worth the expense of stocking herring and kelp. So the burger chain makes the sensible choice to let those outliers dine elsewhere.

Personalizing a U.S. school's education program runs into the same problem, exacerbated by one educational fact of life: public schools aren't supposed to discard the outliers.

Personalized education software must involve the creation of a vast library of exercises and instruction to suit every possible student. Just to teach grammar and usage alone, you would need a vast library of exercises to cover every single rule, every single problem one can have with the rules, and then multiple versions of each exercise geared to the student's reading level and areas of interest.

Our personalized program could get by with a smaller library if we just play to the large middle ground. Only certain rules are problems for most students, most of those students will be reading about the same grade level, and a focus on just three or four areas of interest will engage most students. We can create a relatively small library of materials and hit most students.

But the mission in U.S. public education is not to educate most students. The goal is to educate all students. That means a true personalized education program can't cut corners. A public school is not a burger joint that can say, "Well, if we don't have what they need, they can just go somewhere else." Public schools must serve the outliers, too.

The more you scale up a personalized learning program, the more outliers you have-- and each one will be completely different. That means to truly scale up a personalized education program, you must create a vast library of materials to be prepared for every possible student who could appear-- even if some of those outlier materials might not be used more than once or twice or never in ten years.

So far, most personalized education programs have dealt with this problem by not dealing with this problem and instead creating a program for most students. Some are not really personalizing education at all; instead, they personalize pace. In other words, all students cover the exact same material, in the exact same sequence, but the program lets them each travel down the exact same path at their own speed.

But if personalized education programs have trouble with scaling and outliers, how can classroom teachers possibly cope?

For the most part, easily. The advantage that a classroom teacher has is that she gets to meet the students first. Someone creating educational software right now is doing it for some students who aren't even in school yet, so the programmer must guess and prepare for unknown possibilities in the future, thereby either wasting resources or failing to meet certain needs. The classroom teacher, on the other hand, knows the students, collects daily information about the students, and develops relationships with the students, all of which makes it possible for a teacher to personalize instruction far more effectively than a software program ever can.

This is one reason relationships matter in a classroom. Can the system break down? Sure-- put fifty students in one classroom and that will strain a teacher's ability to maintain the relationships that drive education (that's one reason class size matters). But for true personalized education, you need two persons to form that teacher-student relationship. It's nearly impossible to personalize instruction with just one person involved.

Originally posted at Forbes