I’ve been talking a lot over the last year about the importance of failure in learning, and how play can provide a context for that failure (summarized in this Conversation piece and this article in the Guardian by Andy Walsh).
I have been thinking about different types of failures for a workshop I’m running tomorrow, and considering what makes some failures more useful for learning than others. I’ve come up with an initial mapping of educational failure types based on two scales: how big the impact of the failure is, and how much feedback is provided.
In this model, a trivial fail is one with low impact and low feedback, such as forgetting to signal when driving a car (assuming you don’t cause a crash in the process) – it’s possible to make trivial fails and not notice. On the other end of the spectrum a serious fail has large consequences but also high feedback. For learning, I think that the most useful form of failure is the micro fail with low impact and high feedback – which provides opportunities to learn from the failure and practice in a safe environment. Opposite that is what I’m calling the critical fail, which has high impact but low feedback, such as, say, failing an A-level examination.
How can we create learning opportunities that are more about the micro-fail and less about the critical fail?
These are just initial thoughts but I’m interested in any comments or feedback on this model.
Nic, for me there is work to do on the mapping in the link between Learning and impact.
We learn from both aspects (High and Low impact) . Useful learning also occurs in situations where external pressure is most in your mapping the “high impact” when impact is low it is easy to overlook the lessons as you say “no one notices” in games ,as you know where we talk about low consequences of failure but I am beginning to believe we learn more from high consequential failure (provided we can rectify it )
in terms of impact we learn as you suggest when there is high feedback more than low .
In short I like the ideas and they are obviously in an early stage of development but ‘m not quite with you in your thinking yet.
Paul
Hi Paul
Thanks for the feedback. I suppose what I’m getting at with ‘impact’ is the idea of long term consequence, so that a high-consequence failure that can be rectified would, by my definition (which I appreciate is only in my head at present), be a low-impact failure.
Maybe I need to make a distinction between immediate consequence and long term impact.
Nic
Hi Nic,
You might be right . My thinking comes from personal experience.
As you know I was in the military . I had been instructed to ensure mechanical tools were always in our sea king helicopter and I religiously went through a routing pre-flight to ensure that they were. The one time I didn’t (due to pressures of being scrambled) we landed on the falklands my helicopter broke down which landed us in life or death situation with Argentinians shooting charging up a hill at us. High feedback High consequences suffice to say I never forgot my preflight inspections again learning indelibly form a high impact ,high feedback situation (yes consequences could have been even higher) but The question, does the higher the impact the more focussed we are on the learning and improve the learning (only positing a question I don’t have a cause and effect answer ). Do we focus more on high impact and consequently learn more (ditto previous) …
Paul
This is really interesting to me. Especially as I am currently trying to convince my first year programming students that sometimes failing is the right thing to do. It’s a hard sell for them to get their heads round.
This 2D model seems very elegant, but I have an urge to ruin it by considering some other dimensions. Frequency and timeliness of feedback – I imagine the Micro Fail section would ideally involve small amounts of regular feedback. Motivation and impact – I imagine that it would be difficult to motivate students to perform low impact tasks. Also where is the learning?
I also wonder how learning outcomes align with feedback and/or impact. In my experience some types of feedback are very misleading, but perhaps these often have the highest impact. For example if a first year student scrapes a pass in a module, in the high impact binary world of pass/fail they are doing fine, whereas in reality they should be concerned if they have achieved the necessary learning to enable them to progress further. So now you need an equivalent diagram for students who pass.
So, now I have ruined your lovely 2D model 😉 all that remains is for me to express my profound disappointment that there is no category marked “Epic Fail”.
Well you do just have to go and make it messy and complicated, don’t you? But a multi-dimensional model would allow inclusion of not just the Epic Fail, but the Awesome Fail and the Uberfail.
I wonder whether coding is one example of an activity where failure is inevitable. You code, you test, you fail, you refine, you test, you fail, etc. Or maybe there’s a proper way to do it and I’m just too much of a hacker…
I guess another point is that failure is relative – for some people a scraped pass is a success, for others it’s a failure.