Thanks to early reader Swizec and Harley Lohs for early reading and editing!
I've been reading this book, Accelerated Expertise by By Robert R. Hoffman, Paul Ward, Paul J. Feltovich, Lia DiBello, Stephen M. Fiore, and Dee H. Andrews [1] that tries to explain what the current state-of-the-art is, for not just producing experts faster, but how to reduce the need for re-training when someone has been away for a while.
They tackle things like:
- Transfer: Are there things that help learning transfer from one domain or concept to another?
- Novices vs Experts: What things help/hurt people who are brand new to a subject vs those with some experience? It turns out these tend to conflict and what helps one hurts the other.
- Experts seek out corrective feedback.
- Training can be too easy, some difficulty is desirable.
I'll go over a few of these, highlighting the most important bits along the way.
There are a lot of useful ideas and concepts in the book and I generally like the authors' other works (being NDM people), but it's hard for me to recommend this book for others to read also.
It's a great snapshot summary of where much of the research is for learning in different domains and skill levels. But that's what makes it such a poor fit to answer what, I think, are the more common questions; things like "How do I help a mid-level engineer reach 'Senior' level?" or "How do I help new folks better respond to incidents?"
It is an academic book, and while it provides lots of directions for further reading, the bulk of the "answers" you're likely to walk away with are (accurately) "It depends."
Unlearning is Important
There's this concept they cover called "knowledge shields." Essentially, this is where you take some knowledge or understanding that you have and get stuck in it, using various techniques to "shield" you or block you from learning more.
One of the most interesting points that I got from the book is that developing/increasing expertise involves not just learning, but perhaps more importantly "unlearning."
This is true at every level but becomes more difficult as you gain more experience and expertise.
Unlearning and re-developing new mental models is critical.
Unlearning is especially important because of the "reductive tendency." This is when someone develops some overly simplistic understanding. This is part of how people learn, you have to start somewhere when learning something new and then build on that. Everyone is simplifying some part of their understanding somewhere, including experts.
Knowledge shields can develop and come into play when a learner does something active when they're confronted with evidence that goes against their understanding. One example of this is the "demean effect."[2] This is where a learner might acknowledge something but dismisses its importance, perhaps saying something like "That may be true, but it's no big deal."
There are 23 different knowledge shields identified thus far, though I won’t be getting into them for this review.
What Helps Some Can Also Hurt Others
The current level of expertise of the person learning is important in developing training to help them.
Instructional techniques that are highly effective with inexperienced learners can lose their effectiveness and even have negative consequences when used with more experienced learners. Conversely, instructional methods that work for more advanced learners may be ineffective for less advanced learners.
Transfer
An important idea for the book is that of "transfer," which they define as "the ability to use knowledge flexibly and effectively across application areas".
This is pretty important for training, it basically addresses the question "How much does experience in other things help with the thing I want someone to learn?"
This is true regardless of whether that experience with the "other thing" is a different skill or even a "lighter" form (e.g. a simulation or simplified practice).
I think lots of folks are generally aware of this, it's why gamedays can be helpful, etc. Transfer is why working in ambulances and ERs helped me do incident response in tech. They're not the same, but there was enough transfer that I benefitted.
Transfer can be applied in a few different ways, "transfer across situations" and "transfer across responsibilities". How closely (or not) the training resembles the actual work, from a cognitive perspective, is the "transfer distance". This isn't a literal distance that you can measure, just a notional distance that you may be able to assess or observe.
Ideally, you want to minimize transfer distance most of the time. This is where sayings like "train as you fight," come from. This is an attempt to reduce the transfer distance.
Note that higher fidelity training does not necessarily decrease the transfer distance.
When the relation of the training and work context is apparent, learners can see the value of what they are trying to learn because that value is apparent in the work they are actually doing, and the transfer (of knowledge) conditions and application criteria are likewise relatively salient
Considering transfer can help us design more effective training and practice sessions.
Desirable Difficulty
At first, it may make sense to provide learning situations where people won't make any mistakes. However, adding a degree of challenge in these situations can actually assist with long-term learning. This concept is called desirable difficulty. You don't want the learning situations to be so difficult that the learner gets frustrated or overwhelmed, but you don't want them to be so simple that the learner is bored.
Cases or problems that make initial learning difficult (e.g., learning a sequence in mixed parts, rather than just sequentially), can aid long-term learning and higher order learning
The problem is that it can be difficult to identify desirable vs undesirable difficulties. Among other things, it depends on the learner's current skill.
Automation Requires Greater Expertise
The authors address the common argument, that really it's this increased complexity everywhere that's the problem. As if complexity were dust or debris that we've accidentally allowed into our system. "Who left this complexity here? That's your problem!"
When, in fact, increased complexity often comes from success in our systems and organizations.
They also address attempts at using "automation" as a way of reducing that complexity.
The increased complexity of systems causes errors and poor performance. Therefore, the argument goes, we add more automation to make operations simpler, more easily trained, and trainable to lower levels of expertise.
But, that's not how automation actually ends up working in complex systems.
However, all our lessons learned in cognitive systems engineering and complexity theory imply that more automation may result in simplifications, but they will hide deeper complexity, and at those moments where resilience and high-performance are needed, there will have to be greater operator expertise.
Training for Complex Tasks is Itself a Complex Task
The final thing I wish to leave you with in this review and guide is this reminder:
Training for complex tasks is itself a complex task, and most principles for good instruction are contextual, not universal
This often means the best thing you can do, as it so often is in our work, is to get out there and see how the people you serve actually work.
References
Subscribe to Resilience Roundup
Subscribe to the newsletter.