Sharing Nicely

Failure drives learning. Or rather, debugging drives learning. Not getting it right the first time, making strategic changes, and observing their outcome lets us learn. This is all second nature to engineers, artists, and poets, but it is the complete opposite of how most of education works.

The path is more important than the destination. If learners are told that getting it right is the only outcome that counts, they won’t experiment. And they won’t learn. It is worth redefining the destination of learning and trying to explain what we hope learners will take away from a course. But it is more important to create the space for creativity, exploration, and collaboration that allows them to do things we didn’t expect.

What would a fail-safe learning environment look like?

Stakes are high, but not too high. Learners need to have a stake in their learning. Having a stake means getting frustrated when things go wrong. But the stakes should never be too high, failure in learning should never be catastrophic. It’s important to know when something didn’t work. Having to face mistakes can be frustrating. But it’s not frustrating if each mistake is seen as a small step towards success.

Tinkering encouraged. There are no “no U turn” signs. Good problems don’t have one right answer, but multiple solutions. Fail-safe environments celebrate experimentation as long as it is reflective. Getting it right is secondary. Understanding what went wrong is key.

Fast useful feedback. Feedback needs to be fast, so that we know when things are going wrong while we are doing them. And it needs to be useful so that we can identify different, more successful strategies for the next experiment. How do learners know if they are making progress if there is no feedback and opportunity to revise, improve, iterate?

And here is the crux with MOOCs. MOOCs are too big to support individual constructive failures. The moments that my colleague Natalie refers to as “flopportunities”. MOOCs require a focus on “right” answers, because dealing with ambiguity is hard at the 10,000 user level (let alone the 100,000 user level). This is why in our Mechanical MOOC experiment we added small groups and hope that they provide a safer space to ask questions and get things wrong. And who celebrate effort and experiments over final results.

And the same is true at the institutional level. Many of the early stage MOOC projects build sophisticated platforms, invest large amounts of money in content development, and are basically trying to take over the world. That approach doesn’t leave much space for experimentation, for failing constructively, and for letting more people participate in finding lots of different good models.

In a culture obsessed with “winning” and “success” it can be hard to accept that failure is what drives learning. But as Shimon Schocken said in his TED talk, “grading takes away all of the fun from failing … and a huge part of education is failing”.