Lecture 4: Generalisation and Maintenance Flashcards
When would you avoid using a modelled prompt?
When students do not have the skills of attending and imitation.
_________ is used in fading physical prompts. The teacher begins with as much physical assistance as necessary and gradually reduces pressure.
Graduated guidance
When using a ________, rather than presenting the prompt immediately, the teacher waits, thus allowing the student to respond before prompting.
Time delay
___________ start with the most powerful prompt available. When the target behaviour occurs reliably, move to the next less intrusive prompt (decreased assistance prompting).
Most-to-least prompts
Four Stages of Learning
- Acquisition (learn new skill
and independently responds
correctly) - Fluency (independently
responds consistently and
correctly for most trials) - Maintenance
- Generalisation
*Maintenance and
generalisation should be
planned for in the acquisition
phase. The way to do this is
learner dependent, there is no
one right way to do it.
Generalisation definition
Generalisation = A behaviour change that has not been taught directly. (sometimes called transfer of training)
Three ways to generalise or transfer learning:
o Maintenance (resistance to
extinction, behavioural
persistence, or durability) –
over time; behaviour change
that persists over time.
o Stimulus generalisation
(transfer of training) – across
environments, people,
materials, or conditions.
o Response generalisation
(concomitant behaviour
change) – across behaviours;
the development of related
behaviours not directly
trained.
Stimulus Generalisation
o Stimulus generalisation occurs
when students perform
behaviours in settings or in the
presence of stimuli other than
in which an intervention took
place.
o Child is taught to perform a
skill in one environment but
then uses the skill in another
environment without being
taught in the second
environment.
o Different stimuli (S), same
response (R)
o These stimuli are bound to
differ in natural environments
o For example, you have taught
a child to label a several
different pictures of a
dinosaurs. Prompts are faded
until the child independently
provides the correct response.
After training, you test to see if
the child can label other
pictures of dinosaurs. When
shown a new picture of a
dinosaur, he is able to label the
picture correctly each time.
o For example, a child is being
taught to discriminate
between the colours red,
green, and blue. The teacher
lays out three balls identical in
all aspects except colour and
asks the child to touch the red
ball. The child is prompted and
reinforced for correct
responses. Prompts are faded
until the child independently
provides the correct response.
Now the teacher lays out three
balloons and tells the child to
touch the red balloon. Even
though this has never been
prompted or reinforced with
balloons, the child responds
correctly.
o Red balls and red balloons are
members of the same stimulus
class because the share similar
physical properties and
occasion the same response.
o The more similar the
instruction stimulus (ball) is to
the untrained stimulus
(balloon) the more likely
stimulus generalisation is to
occur.
Response Generalisation
o Response generalisation when
a learner emits untrained
responses that are functionally
equivalent to the trained target
behaviour. Change in one
behaviour results in changes in
another similar behaviour.
o Same stimuli, different
response.
o Conditions determine “R”
change.
o For example, a child is being
taught to give social greetings
by saying “hello”. The teacher
asks people to enter the room
and then prompts the child to
say “Hello” and reinforces
successive verbal
approximations until the child
says “Hello” independently
when people enter. One day a
person that the child was
taught to say “Hello” to enters
the room and the child says,
“What Up?” even though that
phrase had never directly been
taught or intentionally
reinforced before.
o Saying “Hello” and “What Up?”
are members of the same
response class because the
topographies are different but
the effect on the environment
(function) is the same.
o Response continues to occur
even when [external]
reinforcers are not delivered; it
is resistant to extinction.
o Stimulus (S) – Response (R) –
Positive Reinforcement (SR+)
Maintenance
o Maintenance is generalisation
over time.
o The extent to which the
learner continues to perform
the target behaviour after a
portion or all of the
intervention has been
terminated.
o Also Known As: Resistance to
Extinction, Durability,
Behavioural Persistence, and
Response Maintenance.
o For example, a child is
misbehaving in the park. His
therapist implements an
intervention that reduces
problem behaviour to zero.
One year later data is collected
again and problem behaviour
is still at zero.
An Implicit Technology of Generalization (Stokes & Baer, 1977)
o What is generalisation and
how have people programmed
it in the literature? These can
be used in combination and
some are more effective for
specific behaviours or learners.
- Train and Hope
o No actual plan, very common
classroom approach,
generalisation is considered a
passive phenomenon.
o You did not programme it and
just hoped it would happen. - Sequential Modification
o Teach in every context the
behaviour is expected to
occur. If the behaviour should
occur at school, home, the
playground, and grandma’s
house then teach it using the
stimuli and settings relevant
for each behaviour one at a
time.
o Sequentially modifying
behaviour. - Introduce to Natural
Maintaining Contingencies
o Fade SR to the natural
schedule and form expected to
found in the criterion
environment.
o Place behaviour under the
control of natural reinforcers
to increase the frequency of
appropriate behaviours in
DRA. - Train Sufficient Exemplars
o Specifically plan to use
multiple examples of many
different stimuli and settings
when teaching. - Train Loosely
o Do not attempt to program for
specific stimuli and specific
responses.
o Teach across a messy array of
stimuli which is not
appropriate for all students. - Use Indiscriminable
Contingencies
o Learner does not know what
setting or person is going to be
associated with reinforcement
next, similar to an intermittent
SR schedule.
o Teach skill on intermittent
schedule of reinforcement to
teach the student to keep
engaging in the behaviour
without knowing when, where,
and who it will be reinforced. - Program Common Stimuli
o Make the teaching setting as
similar to the criterion setting
as possible.
o Simulate the natural
environment you want the
behaviour to occur in. - Mediate Generalisation
o Something that moves with
the learner across
environments or contexts and
acts as a prompt (reminder) to
engage in the behaviour.
(Language, self-management
strategies, SGD, etc).
o For example, when using FCT
or self-management. - Train to Generalize
o Teach generalisation directly
with specific reinforcement.
Example:
o Teaching a student to raise
their hand insstead of calling
out in class during a group
activity. Use sequential
modification to teach them to
raise hand during multiple
group activities and people
(i.e., seat time, sports, etc.).
Use multiple exemplars on
thick scedule of reinforcement
and then thin it to place under
the control of natural
reinforcers.
o Greeting with sequential
modification.
Categories Related to a Technology of Generalisation
o Natural Maintaining
Contingencies:
Uses trapping manipulations,
in which responses are
introduced to natural
reinforcement communities,
which refine and maintian
target skills without additional
interventions.
This may not be enough for
some learners (i.e., need
smaller steps, massed
practices, and artifical
reinforcers) when first training
generalisation in students with
severe disabilities.
o Training Sufficient Exemplars:
Involves learning sufficient
exemplars of stimulus
conditions or responses.
o Training Loosely:
Is conducted with relatively
little control over the
antecedents and behaviours
involved. Some learners prefer
defined parameters.
o Indiscriminable Contingencies:
Deliberately makes antecedent
conditions and contingencies
less predictable so that it
becomes difficult to
discriminate reinforcement
occasions from non-
reinforcement occasions.
o Common Stimuli:
Incorporates into training
settings social and physical
stimuli that appear in
generalisation settings.
o Mediated Generalisation:
Establishes a response as part
of new learning that can be
used for other problems and
in other situations.
o Training to Genersalise:
Reinforces generalisation as if
it were the target behaviour.
Issues with Generalisations
o Generalizing Teacher
Repertoire (being data driven;
comprehensive and
unrestrictive)
o Using Functional
Contingencies (most
appropriate way to make
lasting change; most
important; extinction probes;
thinning schedules of
reinforcement without
problem behaviour
reoccurring)
o Providing Sufficient Detail (a
plan that specifics
generalisation plan in enough
detail to be reproducible;
observable and measurable;
how long will data be collected
for? In what setting will it be
taught)
o Conducting Outcome
Assessment (are the behaviour
change outcomes socially
valid? Clinically meaningful. In
natural or contrived
situations?)
Generalisation Problem
o When two stimuli have many
similar characteristics, but are
members of different stimulus
classes a child may make this
type of error: Stimulus
Overgeneralisation.
o A child is being taught to tact
(verbally label) pictures of
animals. The child is first
shown a picture of a dog and is
prompted and reinforced for
saying “dog”. Eventually the
child says dog without
prompting. The child is then
shown a picture of a cat for the
first time. The child
immediately says, “dog”.
o Choosing the right stimuli for
multiple exemplar training is
important.
o Now the child is walking down
a crowded street full of
strangers and says, “What Up?”
to every person she sees. This
is may be an example of:
Response Overgeneralisation.
o What should you do about it?:
Teach the child under what
circumstances the behaviour is
appropriate Discrimination
Training (discrimination
training used to overcome
problems with generalisation).
Basic Strategies to Promote Generalisation
o When possible, teach
behaviours in the natural
setting where they should
occur (training in naturalistic
settings is not always possible
for all behaviours).
o Train in a variety of settings.
o Employ a variety of people for
training.
o Use the reinforcers that will be
available in the natural
environment (may include
transitioning from contrived to
natural reinforcers)
o Fade SR schedules to match
the natural environment.
o Reinforce every instance of
generalisation.
o The more the training
environment matches the
natural environment the more
likely acquired skills are to
generalize and maintain.
“Practice how you want play” –
Some Coach I had in Jr. High
If we probe for generalisation and it doesn’t occur (i.e., prompt for generalisation to see if it will occur) then we are not ready to close the intervention and more training is needed before moving onto maintenance.
