Prototypes

(Simulation of Posner and Keele)

Estimated time to complete lab: 50 minutes

Background
Part of cognitive psychology explores the concept of concepts. What cognitive events happen
when you think about a chair? How is the concept of chair represented in the cognitive system?
This is a subtle issue. For example, surely a seat at a formal dining table is a chair, but what
about a recliner, a stool, a couch, or a tree stump? The issue is important because the
representation of concepts is the basis of everything else we can do mentally with concepts. In a
very real sense, how we think and what we can learn is largely determined by how we represent
concepts.

An efficient way to represent concepts would be to keep only the critical properties of a concept.
This set of critical properties is sometimes called a prototype or schema. The idea of prototypes
is that a person has a mental construct that identifies typical characteristics of various categories.
When a person encounters a new object, he/she compares it to the prototypes in memory. If it
matches the prototype for a chair well enough, the new object will be classified and treated as a
chair. This approach allows new objects to be interpreted on the basis of previously learned
information. It is a powerful approach because you do not need to store all previously seen chairs
in long-term memory. Instead, only the prototype needs to be kept.

This experiment allows you to participate in a type of experiment that is often used to investigate
the creation and storage of concepts. It is a variation of a method used by Posner and Keele
(1968), which is one of the earliest studies to systematically explore concept representation in a
controlled way. Rather than using an already well-known concept such as a chair, Posner and
Keele had participants learn patterns of dots. The patterns were variations of a few prototypes,
but the prototypes themselves were not seen during a training phase. During the training phase,
participants learned to classify the variations, with the underlying prototype being the basis for
correct classifications.

After learning to classify the variants, participants were shown a variety of dot patterns in the test
phase. In particular, they were shown three types of patterns. First, they saw some patterns that
had been shown during the training phase. Second, they saw some new patterns that had not been
seen before, but which were variants of patterns they had seen. Third, they saw the prototypes of
the patterns; until this point in the study, the prototypes had never been shown. Classification and
response time performance were nearly equal for the previously seen variants and the prototypes.
Performance was slightly worse for the new variants. This is significant because both the new
variants and the prototypes were never seen during testing. To classify dot patterns that were not
previously seen, the participants must be using a mental concept of what corresponds to the
different categories. Since performance is better for the prototype patterns than for the new
variants, the mental concept is similar to the prototype patterns. The conclusion seemed to be
that people created a mental representation that was a mixture of the variant patterns used during
training (that is, a prototype).

Posner and Keele's experiment led to an intense investigation of concept formation and
representation. Much of that research is consistent with prototype theories. However, there are
aspects of the experimental data (even in Posner and Keele's experiment) that suggest that the
prototype theories cannot be the sole basis for concept representation. For example, our behavior
is often influenced by the properties of individual experiences, and some theories of concept
formation suggest that this alone can account for the data purported to imply prototypes.
Nevertheless, prototypes are a part of many theories of cognition in a variety of forms. This
demonstration is a variation of Posner and Keele's experiment.

Instructions
If you have logged in, you'll see a black rectangle below. Make sure that you can see the full area
before you begin the lab.

Each trial will begin with a fixation point displayed for one second. It will then be replaced by
randomly placed dots. Your task is to classify the dot pattern as A or as B. On the first few trials
you will not be able to properly classify the patterns and so you will have to guess. However, you
will receive feedback on each trial and this way you can learn which pattern corresponds to A
and which to B. You should make your responses as quickly as possible, but try to be accurate.

The experiment consists of separate training and testing phases. The training phase includes at
least 60 trials (trials in which a mistake is made are repeated later in the experiment) and the
testing phase contains at least 16 trials. Each dot pattern in the training phase is a variation of one
of two fixed prototype random dot patterns. The variations are made by randomly taking ten of
the twenty-five dots in a prototype and moving them to a new position.

After completing the training phase, a new set of dot patterns is presented. The transition from
training to testing is fairly seamless, and you may not notice when it happens. The dot patterns in
the testing phase are of four types. One is the prototype that corresponds to the A category.
Another is the prototype that corresponds to the B category. The other two patterns are new
variations of these prototypes (one variation for each prototype). Each dot pattern is presented
four times in random order.

Throughout this experiment your task is always the same. As quickly as possible classify the
pattern as A or B. Press the space bar to start the next trial.

At the end of the experiment, you will be asked if you want to save your data to a set of global
data. After you answer the question, a new Web page window will appear that includes a
debriefing, your data, your group's data, and the global data.

Tablet Specific Details

If you are using a tablet, tap the Start Next Trial button to start a trial. If you think the pattern of
dots is Pattern A, tap the A button as quickly as you can. If you think the pattern of dots is
Pattern B, tap the B button as quickly as you can.

Computer Specific Details

If you are using a computer, place your left index finger on the z-key. Press the n-key with your
right index finger to start a trial. As soon as you've pressed the key, slide your right index finger
over to the m-key. If you think the pattern of dots is Pattern A, press the z-key as quickly as you
can. If you think the pattern of dots is Pattern B, press the m-key as quickly as you can.

What methods did we employ in this experiment?

On each trial you were shown a dot pattern and were asked to classify it as belonging to
Category A or Category B. You were asked to respond as quickly and as accurately as you could.
In the training phase, each dot pattern was a variation of one of two fixed prototype random dot
patterns. The variations are made by randomly taking ten of the twenty-five dots in a prototype
and moving them to a new position. In the test phase, a new set of dot patterns was presented.
The dot patterns in the testing phase were of four types. One was the prototype that corresponds
to the A category. Another was the prototype that corresponds to the B category. The other two
patterns were new variations of these prototypes (one variation for each prototype).

The independent variable is the type of item shown in the test phase, either the prototype or one
of the new variations. The dependent variable is the speed with which the patterns were correctly
classified.

What do we predict participants will do? Why?

You should find that the response time to the prototype patterns is faster than the response time
to the new variants.

How robust is this effect? Are there limits to this effect?

The effect is quite robust and can be substantially improved by inserting more experience in the
training phase.