Mirjam Neelen & Paul A. Kirschner
In a busy online room with around 200 people, responding to all questions unfortunately wasn’t possible. In this blog, we’re taking the opportunity to take some comments and questions and build on them.
One of the building blocks dealt with complex learning (we’ve blogged about this here and here). A fairly ‘simple’ definition provided by Jeroen van Merriënboer (whom we interviewed for our book) is that complex learning entails integration of declarative learning (knowledge-based), procedural learning (skills-based), and affective learning (attitudes). In the medical domain, for example, you need to combine all three to ensure effective learning transfer to the job. This is the most effective approach for many, if not most, professions.
In her webinar, Mirjam used a ‘ladder safety’ example (inspired by an example in Patti Shank’s book ‘Practice and Feedback for Deeper Learning’ (2018)) to explain the difference between a holistic and atomistic design. Holistic design always presents learners with whole and meaningful tasks, right from the start, while atomistic design (the type of design we most often see in organizational learning) breaks a whole task into a set of distinct learning objectives.
She started with an ‘atomistic design’ example (see image). The design example shows fragmentation as well, meaning that the learning activities are mapped back to each objective separately and hence taught piece-by-piece without paying attention to the relationship between pieces (Van Merriënboer & Kirschner, 2018). The participants in the webinar were encouraged to ignore the specific types of activities listed (these were just for illustrative purposes) and focus on the question: “Is a design where learning activities are mapped back to each objective separately effective or ineffective and why?”
Atomistic design example, including fragmentation
A discussion followed, and it was clear that people looked at this question from different perspectives (e.g., “effective, because the objectives are measurable” or “ineffective, because it’s not clear how the objectives will be assessed”). Some participants also got side tracked by all kinds of things such as the visual representation of the table or the emotional involvement of learners (hmmmm).
One participant said: “Ideally, the activity should cover multiple objectives like in real life”. Bingo. The point here indeed was that, if we want people to be able to transfer what they’ve learned to the job situation, the design as shown in the image is not the way to go, as training each objective separately is ineffective for performance because in real life all these objectives are all integrated in one job task.
For example, a job task could be ‘washing windows on the second floor of a building on a windy day’.
In that case, all the objectives on the left-hand side are integrated in that ONE TASK. One participant in the webinar asked “All that was missing was the overarching objective that put the sub-objectives in context. Have I missed something?”
Well, yes 😊. An overarching objective would have made it possible to focus on the actual job task, but that’s not the point. The point is that you shouldn’t treat the ‘sub-objectives’ as separate entities. They should always be trained in the context of the overall learning (job) task and thus, in relation to each other. Hence, you would need to take a ‘whole task (holistic!) design’ approach.
What could that look like?
From atomistic design to a whole task approach
At a very high level, it could look like this:
Whole-task design, focusing on how the knowledge, skills, and attitudes are integrated in relation to the actual task
So, the training would be designed around the whole task and NOT around each separate objective.
Note that the demonstrated task is not a task you would start with. It’s a more ‘advanced’ task. You, of course, would begin with the least complex, yet whole, task first (e.g., washing windows on the ground floor with a stepladder, on a broad level sidewalk with few passers-by and no wind. You add complexity, such as the second floor, no sidewalk, wind, and so forth later).
Complex learning: Integrated Knowledge and Skills
In addition to designing in an atomistic way, another typical mistake when designing learning solutions for organisations is that we distinguish between knowledge, skills, and attitudes and design for each separately. We often design knowledge-based trainings for novices (saying “People just need to understand X” or – worse – “People just need to be aware of Y”) and then only when people become more advanced, do we start focusing on skills and attitudes and we design activities to let them practice them.
In the context of complex learning, this doesn’t make sense. Knowledge and skills (and, if applicable, attitudes as well) always go hand in hand so teaching them separately AND assuming that knowledge comes BEFORE skill is silly.
Some participants didn’t buy it.
The biggest pushback was around the claim that in complex learning, knowledge, and skills go hand in hand (are integrated). People seemed to interpret that as in “You always train knowledge and skills at the same time”, which is not what it means. Participants responded:
“We definitely need knowledge up front in some cases.”
“I disagree that knowledge and skills go hand in hand, you wouldn’t just give a rifle to an inexperienced practitioner and expect them to muddle their way through using it without some prior knowledge?”
Of course you usually need prior knowledge before you can complete a complex task (and understand why you’re doing what you’re doing and why it needs to be done that way). However, in the context of complex learning, you’d only ‘teach’ the knowledge component of a task in the context of the actual task. You’d provide learners, for example, with a modelling example or a worked-out example (see our blog here) in which you present the necessary supportive information in a just-in-time way. In the context of the ‘washing windows’ task above, this worked-out example would explain:
- Which different ladders are available for which situations.
- How you decide which ladder to use for this particular two-story building and why the selected ladder is best fit for purpose.
- How you would inspect the ladder for this job and why that way.
- What you look for when using the ladder and what things are easily overlooked, for example.
- How you recognise hazardous situations when inspecting and using the ladder while carrying up the bucket with water, washing the windows, being very windy, etcetera, and how to make those situations safe.
Of course, because Mirjam used a more advanced task as an example, in reality learners would already have some prior knowledge around these questions at this point, because they would already have studied some worked examples of the simplest form of a ‘ladder task’ and will have incrementally built up their knowledge while studying examples, practising (from highly scaffolded to independent), and receiving feedback. The main point is: You don’t just teach all the different types of ladders, then teach when to use which ladder, then how to inspect ladders, then how to use them, and so forth. Anything you teach, you teach integrated in actual context of the authentic task at hand. This is a huge difference from how we generally design learning solutions for the workplace.
What about the rifle example? Let’s say the whole task is something like: Hit an immobile target at a distance of 10 metres with a small-bore rifle shooting from a prone position. The learner can use a sling for additional stability. This is a beginner task.
Of course, as a beginner, you won’t just get the rifle in your hand and off you go. You need prior knowledge about rifle safety and range safety. However, instead of just teaching all the theory about safety first, you teach the safety rules in the context of the whole task.
You start with a worked example, most likely a ‘modelling example’, where you pay explicit attention to all the steps, the decisions made, and the rationale behind those decisions. The safety rules are explained wherever they’re applicable in the process. For example, first you need to make sure that you are wearing the right safety equipment, then you need to check if your rifle is safe to use, etc. (how do you do it, why this way). As the example unfolds, the learner learns how the rifle is loaded, which way is the safest way and why, how you take position and make sure you’re stable, etc. Again, the safety rules are integrated in the steps as applicable in the context of the task. Of course, this example only illustrative. It’s not complete or completely ‘spot on’ (we’re not rifle training experts; if we were doing the actual design we would include at least one rifle expert, one handgun expert, etc.).
After providing the learner with the worked example, there are various options. One is using a series of partially worked examples, where the instructor starts ‘thinking out loud’, going through the safety rules, loading the rifle safely, then handing it over to the learner so that they can complete the next steps. There might be ‘pauses’ at certain points where the instructor takes over and demonstrates, then the learner completes the next step etc.
Also, when it comes to the safety procedures, it would be good to design what’s called ‘part-task practice’, which means that you design additional practice items that focus on safety because these routine aspects need to be over trained to a very high level of automation; so that you do it without having to think about it.
Incrementally, you move from simple to more complex tasks (e.g., from prone position with supportive sling to prone position without supportive sling, then to kneeling with supportive sling etc.). Again, we’re not subject matter experts when it comes to rifle shooting (we’ve used some of what Zatsiorsky & Aktov (1990) discuss in their article to inform ourselves)., but there might be certain rifles that are easier to load than others and a flat terrain might be easier than a hilly one, no wind vs. head- or tailwind vs. crosswind, etc. In summary, you start with a less complex (yet authentic!) task and provide the learner with scaffolding constantly going on to more complex tasks once the simpler one has been mastered without support or guidance.
For example, 1) just in time information, such as a checklist or rules-of-thumb with what to watch for when loading the rifle (assuming this is a routine / recurrent aspect) and 2) supportive information, such as a set of guiding questions (e.g, “What kind of impact can weather have on your aiming and shooting?”) if the aspects are ‘non-routine’ / non-recurrent (varying from situation to situation).
Within each ‘task class’ (whole tasks at the same level of complexity), you fade the support and guidance (i.e., scaffolding) and when the learner can complete new tasks at that level without scaffolding, you can move onto the next task class (i.e., level of complexity).
Hopefully the difference between atomistic design/fragmentation and whole task design (where knowledge, skills, and attitudes are integrated) is now a bit clearer.
Neelen, M., & Kirschner, P. A. (2020). Evidence-informed learning design: Creating training to improve performance. London: Kogan Page Publishers.
Shank, P. (2017). Practice and feedback for deeper learning: 26 evidence-based and easy-to-apply tactics that promote deeper learning and application (Deep Learning Series Book 2). CreateSpace Independent Publishing Platform
Van Merriënboer, J. J. G., & Kirschner, P. A. (2018). Ten steps to complex learning: A systematic approach to four-component instructional design. New York, NY: Routledge.
Zatsiorsky, V. M., & Aktov, A. V. (1990). Biomechanics of highly precise movements: the aiming process in air rifle shooting. Journal of biomechanics, 23, 35-41. Retrieved from https://www.researchgate.net/profile/Vladimir_Zatsiorsky/publication/21073773_Biomechanics_of_highly_precise_movements_The_aiming_process_in_air_rifle_shooting/links/59dcf46c0f7e9b11b6234a3f/Biomechanics-of-highly-precise-movements-The-aiming-process-in-air-rifle-shooting.pdf