Professional Documents
Culture Documents
- One dimension of behaviour that may be shaped is variability itself. When behaviour is
more variable, it provides more variants that may be reinforced and therefore may make
shaping quicker and easier.
- One more dimension of shaping is illustrated by the rat and the counterweighted lever.
Ordinarily in shaping, you watch current behaviour and deliver reinforcers based on
when you see.
o In shaping procedures that set thresholds, your judgements are not which
responses to reinforcer but rather how quickly you can move the threshold up.
o If you move up too quickly and responding stops, you might have to drop the
threshold down pretty far to being back responding before you can start moving it
up again.
Chapter 9
1) In your own words describe how shaping is used to establish new behaviors.
Describe in your own words the key terms used in the process (i.e., differential
reinforcement and successive approximation).
a. Differential reinforcement – reinforcing the behavior you want to have and
not reinforcing the behavior that you don’t want.
b. Shaping is based upon differential reinforcement. At successive stages, some
responses are reinforced, and others aren’t
c. The criteria for differential reinforcement change as responding changes, in
successive approximations to the response being shaped.
4) In your own words describe the response parameter. What is the difference
between the function class and descriptive (nominal) class? Provide examples.
a. Distinction between the behavior that is reinforced and the behavior that is
generated by reinforcement
b. DESCRIPTIVE/NOMINAL – THE BEHAVIOUR THAT IS REINFORCED
c. FUNCTIONAL CLASS – THE BEHAVIOUR GENERATED BY REINFORCEMENT
- What can be shaped has limits and successful shaping sometimes depends on knowing
more than just when to deliver a reinforcer.
- Responding can vary not only in location or force but also in topography or form,
direction and so on.
- Differential reinforcement can be based on any response dimension, so any dimension or
combination of dimensions might provide the defining properties of an operant class.
Mediating Behavior
- When we observe sequences, another issue is the function of the responses we see in the
sequence
- When responses make sequential contingencies more likely to be met even though they
are not required by the contingencies, they are sometimes referred to as instances of
mediated generalizations.
- In research settings, mediating behaviour it typically unplanned and its function is
determined only by experimental analysis.
- Natural environments provide many opportunities for establishing such sequences,
especially where situations require waiting.
- If mediating behaviour during a wait extends the time and therefore changes the
likelihood that some response at the end of the wait will be reinforced, that behaviour
may be properly regarded as a product of the contingencies.
Variable Behaviour
- The close correspondence between the class of responses with consequences and the class
of responses generated by these consequences is the criterion for speaking of an operant
class.
- Our major interest is with the dimensions along with responding conforms to the class of
responses that is reinforced.
- the structure of behaviour is such that we can’t always define such dimensions
independently of reinforcement contingencies.
- Example, if a porpoise’s backflips were reinforced in one session, slapping the water with
its tail reinforced in the next and beaching itself at the side of the pool reinforced in the
next one, after several sessions, the porpoise began to emit responses in each new
sessions, such as leaping up from the water with a corkscrew spins, that the
experimenters had never seen before.
o Novel behaviour must be emitted before it can be incorporated into other
behaviour. But the fact that we have difficulty measuring it doesn’t rule out
novelty or other complex dimensions of behaviour as properties that define
operant classes.
- Another dimension that cannot be defined independently of reinforcement contingencies
is the variability of responding, which has also been a target for differential
reinforcement.
- Study showed that reinforcement could in fact maintain highly variable behaviour and
these results led to a research program that demonstrated that variable responding was a
dimension of behaviour that could be reinforced.
- Novelty and variability can only be properties of responses in the context of other
responses that occurred earlier.
Chapter 10
1) Describe and provide examples of the dimensions in which responses can vary
a. Responding can vary not only in location or force but also in topography or form,
direction etc.
b. Differential reinforcement is based on any response dimension, so any
dimension or combination or dimensions might provide the defining properties
of any operant class.
The significance of these examples is that we must be cautious about taking response rate as a
fundamental measure of the effects of reinforcement.
LECTURE
- Learned helplessness
o When people know they can control their environment they do much better.
o When people perceive they have the control of the situation they do much better
overall
o Offering choice opportunities – perform better – much less problem behaviours
- DR effects – response similar to those that have been reinforced occur more frequently.
Responses similar to those that have NOT been reinforced occur less frequently
(Extinction)
Dimensions of behaviour
- Topography – form of the behaviour
- Frequently
- Latency
- Duration
- Amplitude (magnitude)
Successive Approximations – usually when shaping the child must have some prereq in
repertoire
Generalizations
- Behaviours not being reinforced are generalizations when they continue without
reinforcement
BEHAVIOUR CHAIN
- Sequence of responses in which the result of each response (except for the last) is a
conditioned reinforcer for that response
- E.g. putting on pants – completion of step 1 is an SD for the next step
- 3 types of chaining
o Forward chaining
o Total task chaining
Teacher prompts when necessary
Continue teaching until mastered all steps
o Backward chaining
Puzzle example