Archives for September 2006

Donald Clark, ex-CEO of the Epic Group in the UK, has got a great blog, where he cheerfully skewers misconceptions about learning. In a rant I was pointed to, he takes aim at Gagne’s Nine Dull Commandments. As the subsequent comments point out, the problem isn’t necessarily with Gagne’s elements as it is with their interpretation and a belief in their rigid order.

It’s not a case of “if you follow it (without insight), they will learn”. The typical interpretation of many of the elements is so rote that they counter what they’re supposed to do. I’ve tried to suggest an improved approach in my white paper about the Seven Step Program to eLearning Improvement (NB: PDF), and it’s nice to see others picking up that a mechanistic implementation of Gagne’ (or anyone else) won’t hit the mark.

As I previously mentioned, I got the privilege of co-hosting a day of meetings for the UK’s visiting mission on eLearning. Sponsored by the Department of Trade & Industry, these missions send a small panel of experts from industry and academia to review and report on relevant international activity. They each have specific areas of responsibility, and are to bring back the outcomes that they determine. I have to say that the group seemed very experienced and aware, and I was eagerly awaiting the report.

The report (PDF) is now available to all. This report includes chapters on mLearning, games, performance support, and more. I haven’t yet read the whole thing, but what I have read looks suitably insightful. I recommend having a look at this critical evaluation of eLearning in the US.

Jay Cross’ Informal Learning Blog, InformL pointed me to this blog post about the need for manuals and the problems with them. The diagram says it all, really.

My PhD work included mental models, and the research is robust: that with models, learners can forget some steps and regenerate the missing bits; that models provide the ability to predict what will happen or explain what happened; and troubleshoot. But what we get in instructions are rote procedures to do something, not oriented around our goals.

When I tried to learn Macromedia’s FreeHand, their tutorial had you build a picture. It was useless. It was only by experimentation that I discovered that what I thought were the atomic elements, shapes, were actually paths that could be manipulated, e.g. cut (see the Quinnovation logo). They didn’t provide a conceptual model that talked about paths and how everything, text, shapes, etc, could be translated to them (and had to be for real manipulation). I may still have it wrong, as it’s also known that individuals will infer models, possibly incorrectly, and without guidance can retain those models with great persistence.

I’ve been trying to get most elearning to focus on the underlying models instead of rote procedures (it’s one of my seven steps to better elearning points; warning, PDF).

I also note that the claim about systems being designed to not need manuals doesn’t make sense for anything more complicated than a toaster. Unless, of course, you’re going to give me unlimited bandwidth and resources, but most people want to keep the size of their devices under control and provide the full set of features people want.

I argue that it’s not about training, it’s about supporting performance, and that includes responsibility for a broader picture. Currently too much of this is siloed off in different parts of organizations, so software engineers write help systems, a different group writes manuals, the training group prepaes training, etc., all from the same information (hence the push for ‘single-sourcing’ and content models). We can do a better job if we start from the models, and populate all these forms of information in an integrated, cross-referencing, and encompassing strategy.

If, as the experts have it, customer experience is the new differentiator, having a usable system coupled with coherent support ought to be part of the picture.