Being PhD student is a great thing. On the one hand, you have to specialize in single topic so that when you graduate you become an expert in that topic. On the other hand, you have to know the related work that may be surprisingly broad at times. For example, I’m doing research in software engineering. My goal is to help software developers to be more productive and to produce software of higher quality. Cool. Recently, however, I needed to check out publications from cognitive psychology! Why would I do that? We were interested in how people learn new things and how brains of experts differ from brains of novices. I learned a lot and want to share some of the findings. You can use these findings to be more effective at explaining things to others. This blog post is based on our recent publication.
Base Your Explanations on Examples
Let’s say you want to explain your younger cousin Newton’s law of gravity. Certainly, it’s not enough to give just a bunch of equations and expect the cousin to understand the topic. Why not? Because learning is about building mental models. Mental models represent our intuitive understanding of the world in the form of generalized information, such as rules. So how to build the intuition? You’re right, through examples! Examples capture intuition. You need to see multiple examples of a single thing or phenomenon. For instance, to start understanding gravity you need to see an apple falling from the tree, a sandwich falling on the floor, or you can even jump out of the window. Only then will then brain start noticing commonalities, will generalize information, and will come up with a rule. Of course, these examples do not fully explain the law of gravity, but provide right intuitions.
Present both Examples and Abstractions
OK, we agreed that presenting just examples is not enough. In fact, the best knowledge transfer occurs when learning comprises both examples and abstractions (such as rules). So to fully explain gravity, you’d have to show both examples and the mathematical equations. Then your cousin should be able to relate examples (that represent intuition) to the equations (that represent general rules). Why do we need to present examples and abstractions simultaneously? First, examples typically capture just a small part of a general phenomenon. In many cases you would need a huge number of examples to appropriately model the general rule. Of course working with a huge number of examples is incomprehensible. Second, abstractions capture the knowledge more completely than examples, but are equally difficult to comprehend on their own.
Novices Need Examples, Experts Prefer Abstractions
Your younger cousin is clearly a novice and knows very little about gravity and physics. In his case examples are useful to build intuition before presenting the general rule. Would you use the same approach when talking to a professor of physics? Quite likely they’d get bored with your examples and would consider them trivial. Why? Because the professor is an expert and has all the required intuitions and knowledge. A better approach is to first present a general rule and then use examples only for clarifications. In that case examples compensate for the missing knowledge. In short, novices need examples right-away, while experts need them only for clarification.
Show a Variety of Examples
Now we know that examples are useful. But, think about it. Are all examples equally useful? No! If examples are too similar (such as showing two apples falling from the same tree), they provide little information. If they are too different (such as showing and apple and a solar system), you cannot easily find commonalities and generalize the information. Effective learning requires a specific variety of examples. The most effective are near-miss contrasting examples. Such examples emphasize critical differences and facilitate building flexible abstractions (mental models). We can distinguish positive and negative examples. Positive examples represent something that is correct (for instance, an apple falling from the tree). Negative examples represent something that is incorrect (for instance, an apple attaching itself to the tree). They explicitly show things and phenomena that are disallowed.