via Wall of People Being Awesome on Chat Roulette.

A post on Philipp Schmidt’s stream got me thinking about how we might tap into the 20 years of research on Peer Instruction to better inform peer learning initiatives. 

I’m assuming here that readers are familiar with the Peer Instruction research (if you are not, you really should take a look at it — some of the more amazing results are coming out of that area).

The question is how you take Peer Instruction (which after 20 years of tinkering and research has become a robust method) and transfer it to the web in a way that reduces the need for instructor facilitation.

Here’s one idea:

Consider a piece of software that puts a question to 1000 students in real time scattered around the web. The students answer, and then are automatically paired with people that choose a different answer through chat or web cam. Each person explains their reasoning to one another for a couple of minutes, then people revote. If the questions were well designed you might be able to get some of the effects of PI.

The voting piece turns out to be a big factor, BTW, because it focuses the mind and the discussion on an objective, leading to a more productive exchange (at least for many applications). I find it actually work wonders in my F2F classes.

But I could certainly see how with a little coding you could have, if not instructorless sessions, then at least sessions which were mostly facilitated by software. And the nice thing here is you already have literally 100s of articles that have researched what makes these sorts of experiences more effective and what makes them less effective — you’re not shooting in the dark quite as much.

I could also see a sort of rotation thing built in — if you (with your peer instruction help) are able to answer a certain percentage of the questions correctly (say 95% correct, rolling average over last 50 questions) you graduate. So people roll out, new people roll in, etc.

If I built this, by the way, I would call it some pun-like name based on Chat Roulette. Because I see it as that level of simplicity — you answer a question and suddenly the system sends you to videochat with a random person that disagrees with you, and the 3 minute timer for the discussion begins to count down. You and this person have to figure this out before the vote, right?

How fun is that? 

Would this work? Maybe, maybe not. But at the very least it’s an attempt to bring something with demonstrated success into the web sphere.