Part of learning about anything new involves learning about
the objects and individuals in that arena.
If you are watching football for the first time, before you can
understand the rules, you have to know about the field, the various players,
and objects like helmets and the football itself.
This issue is particularly important in classroom settings.
Part of the science curriculum, for example, teaches students about the way the
world works. If you are going to learn
about oceans, then you need to learn about the kinds of animals that live in
the ocean before you can understand the ecosystem of the ocean.
So, what is the best way to teach people about objects?
This question was taken up in a paper by my colleague Brad
Love and his former student Yasu Sakamoto in the December, 2010 issue of the Journal of Experimental Psychology: Applied.
A traditional way to teach people about new objects (like
the fish in the ocean) is to have them learn to classify different objects. I remember having to do a leaf collection in
7th grade, in order to learn to classify the various trees in my
neighborhood. In classification, you see
the various properties of objects (like sharks, fish, and plants) and you learn
to identify the category that they come from.
There are other ways to try to teach people about things,
though. In a 1998 paper in the Journal of Memory and Language, my
former student Takashi Yamauchi and I introduced a technique called inference learning. When doing inference learning, people are
told what category something belongs to and have to predict some of the
features it might have. You might see a
particular shark and ask how many pups it normally has.
Sakamoto and Love compared learning through classification
to learning through inference in a set of elementary school students. The students were learning about sharks in
the ocean. The students learned about
some sharks by learning to classify them.
In classification, they would see an animation of a shark described by
properties like its size, eating behavior and number of young that it has. After seeing the shark, they would have to
state the type of shark it is.
In inference learning, the students would be told the type
of shark (“This is a Tiger Shark”), and they would see some of its properties
(like its size and eating habits) and had to predict some other property (like
the number of young it normally has).
Across different questions, students would have to predict values of
different properties.
Even though classification is frequently used in classroom
settings, inference learning was much more effective. Students were able to remember more about the
sharks even a week later if they learned about that shark by predicting
properties it has than if they learned about it by classifying it. They were even able to learn some information
about the properties of the sharks that they were never asked about, though
they learned less about these properties than about the ones that were the
focus of learning.
The results of this study fit with a growing understanding
of the way people form categories. The
knowledge that we store is set up to help us act on the world. As a result, we draw on the procedures we
used when learning something to determine what we might need to know about that
item when we encounter it again. If we
just learn to classify items, then the cognitive system assumes that the most
important thing about the object is that we be able to identify it again in the
future. As a result, we learn only
enough about the object to be able to figure out what category it belongs to.
When we have to predict a range of properties about things,
though, then the cognitive system assumes that we need to know a lot about
it. As a result, we seek to understand
the variety of properties that the object has to ensure that we will be able to
make accurate predictions about it in the future.
When we teach about things in school, then, we have to be
careful to get students to interact with the information in ways that will help
them to use that information later.
Asking students to make predictions about aspects of new items is one
powerful way to ensure that they learn about them effectively.
