How People Learn

Designing Education and Training that Works to Improve Performance

Nick Shackleton-Jones

• Éditeur: Kogan Page
• Année: 2019
• ISBN:
  9780749484712
• Édition: 1
• Sujet: Psychology
• Sous-sujet: Industrial & Organizational Psychology

What if we have been wrong about learning? Learning may have more in common with marketing than we thought. Looking at marketing and learning's common root, How People Learn shows L&D professionals a new way of thinking about learning by exploring what happens when we learn. It considers applications from AI, marketing and ethics and is informed by psychology and contemporary neuroscience in order to show L&D professionals how to design training with their employees in mind so that training makes a real difference to skills, capabilities, performance and development, rather than being a waste of time, money and resources.Using the author's '5Di model', How People Learn demonstrates how to define, design and deploy training in a user-centred way so it works both for and with employees. It also includes guidance on what training resources to create when employees are actively searching for learning content. Using this book, L&D practitioners will be able to use pull and push techniques to provide content that people use and experiences that transform their behaviour. From how to use simulations, storytelling and anticipation to the importance of observation and status, this book gives L&D professionals everything they need to build effective training programmes and learning experiences. With a foreword by Dr Roger Schank, the Chairman and CEO of Socratic Arts and Executive Director of Engines for Education, and case studies from companies such as BP and the BBC, this is an urgent read for learning professionals.

04

Learning design – push or pull?

When I left my job as a psychology lecturer it was, in part, because I was growing tired of telling people how they could change the world if only they took the time to apply some of the theory I kept banging on about. I wanted to try it for myself. A few years later I was running a team of developers, and 20 years ago we were creating digital training that would still look futuristic today: dynamically generated ‘AI’ characters that interacted with learners in randomized simulations, virtual environments to explore, and so on. We were both excited by, and proud of, the work we were doing. We had developed an instructional approach we called the See–Hear–Do model (not to be confused with the Scooby-Doo model), based loosely on Jerome Bruner’s ‘Stages of Representation’ theory,1 according to which people encode information in three formats: enactive (by doing), iconic (through images) and symbolic (through language). By ensuring that learners used all three encoding modalities we were greatly enhancing the effectiveness of our learning content. Or so we thought.

I think I imagined it would be nice to appear on the conference circuit, pointing to a bar graph showing how much more effective our approach had proved to be. As luck would have it, we were host to a work placement – a visiting Dutch research student called Marguerita – and together we designed an experiment.

In our experiment the same information – information about the solar system – was developed into five formats. Each format was different: the most basic version was simply plain text, other versions integrated more media – audio, video, culminating in a version incorporating audio, video and interactive exercises. All the modes of representation.

Groups of university students were then given one of these different formats and the same task: use the materials for 30 minutes, then take a test to see how much they could recall. Our hypothesis was that the richer encoding formats would produce superior recall scores. It helped that at the time the received wisdom was also that ‘interactive’ learning was better than cheaper, less interactive formats.

We ran the experiment. We added up the scores. The results were clear: there was no significant difference in the effectiveness of the various formats. In fact, the students who had studied the text-only format performed slightly better, on average. Imagine that: realizing that perhaps an entire industry based around the production of online content could be replaced – with text documents.

Our research wasn’t published or peer reviewed. I could see that there were shortcomings in the size and representativeness of our sample. A university cafeteria was hardly a controlled environment. But I had supervised enough student psychology experiments to know that our research design was fundamentally sound and that if our instructional design approach was even half as effective as we believed it to be, we should have seen an effect.

As it happened, the results also tied into an observation that had been bothering me for a while: the internet was booming and fast achieving recognition in its own right as a learning tool. But the vast majority of information that was being served up was just – well, plain text. And people didn’t seem to mind. In fact, if you were looking to find something out you probably preferred plain text over other formats.

Even as the learning and development industry were busily handing out instructional design accreditations to each other, it was becoming clear that instructional design was not significantly improving learning in the real world. People would actively avoid formats (such as e-learning modules) that incorporated instructional design, and prefer formats (such as the pages served up by Google, or videos served up by YouTube) that had none. Instructional design, I realized, is mumbo-jumbo.

There is, of course, a great deal of research in support of instructional design. But rather like our old friend Ebbinghaus’s forgetting curve, this research is both accurate and misleading.

Many people struggle with the idea that scientific research can be both accurate and misleading – but in reality this is generally true. The history of science shows that one theory is eventually replaced by another, better, theory. It is always the case that the old, incorrect, theory is well supported by research, and usually the new theory has little or no support (at first). In other words, it isn’t hard to find evidence to support an idea that, ultimately, turns out to be wrong. So how does a theory with little or no support replace one with plenty? Through explanatory power. Invariably there are things that the old theory just never quite explained. The new theory accounts for these.

It is surely a legitimate challenge to instructional design that learners generally prefer formats without it to formats with it, when faced with a real-world task. Give me Google over an e-learning module any day. Instructional design research often repeats Ebbinghaus’s mistake: in trying to determine the best learning formats, it overlooks the fact that in the real world we learn about the things that matter to us: how to fix a router, when to plant petunias, how to bake chocolate brownies. In experimental contexts, real-world phenomena are often excluded, with the consequence that the findings reflect minor variations in cognitive processes – and completely miss the things that really matter. Put simply: people tend to find the evidence they are looking for, but may miss the bigger picture. Learning researchers are often like people studying the corner of a dark room with a torch. They can tell you a lot about little bits of fluff, but have yet to notice the table, chairs and chandelier in the middle of the room.

The students in our own experiment learned equally well from each of the formats because they knew they were going to be tested – and were therefore motivated to consume the information, whatever the format. This variable – their concern – vastly outweighed the significance of anything else that we were testing, such as variations in content format. We had given them a reason to care.

So let’s start by laying the foundations for learning design with a learning design model. There are many models for learning design, but only one that reflects the reality of learning as described by the affective context model: the push–pull spectrum.

As I have stated above, the affective context model proposes that we do not, in fact, encode ‘knowledge’ in our memories: instead we encode the pattern of emotional reactions that we have to experiences, and use these to later reconstruct experiences.

The important thing from a learning perspective, however, is that people are not blank slates: what they react to is partly nature, and mainly nurture. Some people will be astonished by architecture that others don’t even notice – moved by classical music that leaves others cold – or left speechless by a work of art that leaves others unmoved. Some people even like Country and Western songs.

Learning design can therefore rest soundly on this conceptual foundation: since our learning is governed by our concerns, which in turn determine whether or not something has affective significance for us, in any context we are either responding to the concerns that a person already has (‘pull’), or we are trying to create new concerns (‘push’).

This means that there are two grand classes of learning design that we can undertake:

1. Pull. Where people care about something, we do not need to provide affective significance. We can simply provide the resources they need. This is why plain text or simple video is preferred by people using Google or YouTube – because there is something they are trying to do. Something they care about. The vast majority of our learning is driven by the things we are trying to do, and which we care about. This is the ‘pull’ condition.
2. Push. Where people do not care about something, we need to provide affective significance. We cannot simply provide information and expect learning to take place. If someone is uninterested in safety, then providing procedural guidance will do little to alter their behaviour. If a student cares neither for history nor for the outcome of the history exam they are soon to take, then providing more textbooks will be of little use. We need to provide an experience, such as a challenge, experience or story, that will add affective significance to a situation. This is the ‘push’ condition.

It should immediately be obvious that these two approaches go hand in hand. If, for example, a young person cares greatly about a particular computer game, they will quite happily consume encyclopedic tomes on game strategy without the need for encouragement. Teenagers concerned about their appearance will demonstrate impressive attention spans as they sit through hours of hair and makeup tutorials in order to achieve the desired result. If we can make people care about something or other, they will use the resources we give them. Hence the learning is not something we ‘do’ to people – we just support it or create the conditions for it to happen. This is why, in an important sense, there are no ‘learners’ – because people are generally just trying to get something done, and learning is a way to do that. Describing people as ‘learners’ is a bit like describing people as ‘breathers’. Learning is a by-product of pursuing the things we care about.

You can probably already imagine how our modern world presents huge challenges for education: on the one hand, students are increasingly saturated in a hyper-charged diet of things they do care about – sex, games, relationships, comedy, action, shock, gossip and social approval; on the other, they are expected to function in a perversely rationalized environment where they are required to regurgitate information that has little of no significance to them. It’s hard to imagine a more radical pattern of psychological stress – rather like feeding someone on a rich diet purely to deepen their suffering during starvation.

In normal life, our interactions with the world and with other people are often somewhere in between ‘push’ and ‘pull’. Take conversation for example: in a good conversation people take a little time at the outset to ‘establish rapport’. What does this mean? It means people try to understand what the other person’s particular concerns are and to find some common ground. For example, two people may start out by remarking on the inclement weather, then a little later on discover a shared passion for food and for local restaurants. Each person ‘pulls’ recommendations from the other. At one point, one person remarks that the fish at a local restaurant is exceptional – to which the other responds that they don’t much like fish. Their conversational partner then shifts into push mode: ‘Oh but you really must try the monkfish – I mean if you haven’t tried it you are missing out on one of the most incredible flavours!’

It is now clear that one person is trying to make the other care about something that they didn’t before – to try something new, and to change the way they feel. If they have established sufficient rapport – if the other person now cares what they think – then this push attempt might be successful. Notice how similar this is starting to look to the way education should be.

A conversation really is a good metaphor for learning, since a conversation is one of the oldest learning methods we have. Conversation is first and foremost a vehicle for stories. If one person has nothing to say that interests us, then it isn’t a good conversation. If a person takes no time to understand us, and simply ‘pushes’ their concerns the whole time, then that person is a ‘bore’ and the conversation will feel, literally, like a lecture. Finally, we cannot enjoy a truly great conversation with someone who is exactly like ourselves – it seems comfortable at first but eventually we crave a different perspective.

In this light we can make sense of the enduring appeal of coaching, which reliably features in the wish list of people asked about their development. Coaching (or more accurately ‘mentoring’) is a format that can be both experience and resource – both push and pull, depending on the circumstance. Sometimes we turn to a coach for help, sometimes they challenge us. They take time ...