When I was a grad student, my advisor looked a lot at error. HIs particular focus was to prevent it through good interface design. He characterized them as of two types: slips and mistakes. Slips are when you have the right intent, but from elements of our architecture end up making the wrong move. Mistakes are when your intentions are wrong. From the learning perspective, it’s the latter we want to address.
So, I’ve argued that there’s some randomness in our architecture. We’re bad at doing rote things, even when we know what to do. Athletes, for instance, practice for hours a day, day after day and their competitions are typically decided by who makes the fewest errors. In fact, we’ve identified and created support tools to minimize these errors, like checklists. We’re supposed to design systems so they minimize the opportunity for slips. These don’t always work; I was embarassed in my most recent presentation to find a couple of typos. They clearly had come in via auto-correct, but I’d missed them subsequently! There may well be one in this post, too, some form of slip or another.
More importantly are what are termed ‘mistakes’. Here we’re starting with the wrong intent. This, from a learning perspective, is what I term a ‘misconception’. Here, we’re supposed to be using a particular mental model to guide our performance, but we might not have the right model, or no model and we import a wrong one. Again, we should design to minimize these as well, but we also want to make sure we’ve got the right mental models to begin with. This happens if we haven’t been given the right one, or activated an irrelevant one inappropriately.
Ideally, we want to identify the most likely ways we go wrong, and then make sure we understand what those are and why. From there, then make them available as options in practice. Then we can remediate them at the moment. Of course, we’re also supposed to have the right model, and highlight how the model guides performance in examples, and then again through the feedback on both correct and incorrect performance.
Designing practice that supports making the right decisions in context, and also the opportunity to make the wrong choice and get useful feedback, is a key learning design skill. Quiz questions that just test knowledge aren’t likely to lead to meaningful difference in performance. Practice where the alternative to the right answer are silly or obvious (unless you really know the model) is a waste of your resources and your learner’s time. Make practice meaningful. Not doing so is an error, too!