DESCRIPTION OF THE STRATEGY

It is often desirable for clients to acquire new ways of responding, that is, to expand their
behavioral repertoire. Shaping is a procedure for teaching children to do new things. When
using the shaping procedure, new behaviors are constructed from the bottom up. The
procedure starts with a response that the child is already able to do, and works step-by-step
from this response to the goal behavior. Each small step along the way is reinforced, and once
a step is accomplished, an additional small requirement is added. This gradual building
process facilitates motivation and progress and continues until the goal is reached.

Technical Description
Shaping involves differentially reinforcing responses that successively approximate a terminal
response class resulting in new and/or more complex and elaborated behaviors. The basic
behavioral principles operating in shaping are reinforcement and extinction, which when used
in combination, such that only a certain class of responses is reinforced while others are not,
is referred to as differential reinforcement. For example, when coloring with crayons, children
must acquire responses of adequate force to produce marks on paper. One set of responses,
those involving sufficient force to produce color, is reinforced, while responses of insufficient
force produce no reinforcement.
In differential reinforcement, the response requirement for reinforcement is generally
relatively stable. For instance, a certain class of forces, once established, will almost always
produce color on paper. How, though, is the response class established? Here is where shaping
enters the picture. Shaping involves a sequence of differential reinforcements where the
provision of, or the amount of, reinforcement is linked to the progression of responding. As
performances that more and more closely resemble the ultimate performance are established,
previous steps in the sequence no longer produce reinforcement (or produce less
reinforcement than current and subsequent steps). To illustrate, imagine teaching a child who
failed to learn to color. First, the child might be reinforced with food for holding or grasping a
crayon. Then, holding the crayon with the tip touching the paper would be required for
reinforcement. Next, some small movement of the crayon on the paper with enough force to
produce faint color would be required. Finally, only movements of the crayon that produced
solid color would be reinforced. Notice in the coloring example the interaction between an
arbitrary reinforcer and a naturally occurring reinforcer. Early responses in the sequence were
reinforced by food; however, as shaping progressed, the natural contingency between using
adequate force and seeing color on the paper could also begin to reinforce appropriate
responding along the force dimension. That is, too much force would result in the crayon
breaking, too little force would produce no, or only faint, color, while a range of forces
produced the clear sight of color.
It is important to note that shaping is involved in the acquisition of many responses but is
often not recognized as such because it is occurring naturally or in the context of the child's
interaction with a social environment that is effective in intuitively shaping behavior. For
instance, when a child first responds to the letter W by saying double cue, caretakers may
provide praise but later will require the child to more closely approximate double you
before praise is given. When the child begins to string letters together and first utters a-b-c-de-c-d-e, the parents may praise the partially correct response. However, soon such praise will
be reserved for only more extended correct utterances. The programmed use of shaping in

clinical settings is simply a formal extension of the process that is occurring in a less formal,
but no less influential, manner across many domains during development.

Clinical Description
When using shaping clinically, five basic steps can be identified. The first two steps occur
before the shaping procedure is actually initiated.
The first step is to conduct a thorough baseline assessment to determine how the problem
relates to a deficit in the content/form, duration/amount, force/ intensity, or latency of some
aspect of the child's behavioral repertoire. In some cases, an apparent deficit is due to a lack
of opportunities to respond or to existing contingencies that reinforce incomplete or weak
responding (e.g., when a caretaker withdraws a demand at the first sign of resistance or
intervenes at the first sign of difficulty). Shaping can still be employed in such situations;
however, altering the antecedent conditions or changing the existing contingencies will
greatly facilitate the process.
Once the problem has been identified as involving a deficient response repertoire, the second
step is to identify a reinforcer (or set of reinforcers) that can be used in shaping the response.
Depending on the clinical situation, one could employ a formal preference assessment, where
across several trials the child chooses between two or more potential reinforcers offered
simultaneously, to determine the reinforcers that are likely to be most effective. Alternatively,
therapists can often make educated guesses based on information from caretakers or their
knowledge of the client. In some cases, trial and error during the shaping process may be
required. In all cases, it is worth remembering that a reinforcer is defined based on its function
(its response contingent immediate presentation increases future responding), and that an
effective reinforcer at one time may not be effective at a later time. When possible, it is
desirable to arrange the shaping procedure so that the ultimate control of the response will be
by naturally occurring or conventional social reinforcers (e.g., praise).
The third step is to define the class of behavior in which the client would ultimately, at the end
of the shaping intervention, be able to engage. When possible, it is useful to be as specific as
possible when identifying what constitutes the terminal response class and what does not.
Finally, when possible, it is important to specify what will define learning of the terminal
behavior. It is important to remember that saying a behavior is now part of a client's repertoire
is merely to say that the client emits the behavior. Speaking of a client's repertoire is an
abstraction; the new behavior does not exist somewhere within the client in a place called the
repertoire. As such, it is not enough to simply observe the terminal behavior once, discontinue
the procedure, and conclude that the performance is now part of the client's repertoire.
Instead, it is important to specify, to the extent possible, what indicates a successful
conclusion for this particular client (e.g., when asked by a teacher or parent, the child can say
the entire alphabet accurately across 10 consecutive occasions, with each occurrence taking
less than 10 seconds to complete).
The fourth step is to identify an initial behavior that the child currently emits with some
frequency and that shares as many features as possible with the terminal response.
Reinforcement occurs after a response and increases the future likelihood of that response
class. If a relevant response is never emitted, reinforcement cannot be received. Thus, the
initial response chosen must be maintained currently at some minimal frequency. That said,
one need work backward from the terminal behavior only far enough to find a relevant

response that is already being emitted prior to implementing the shaping procedure. However,
it is possible that you will need to begin the procedure with an extremely basic response class
such as simply orienting, vocalizing, or touching. Once the initial behavior is identified and its
frequency increased, it is then time to increase the threshold for reinforcement by requiring
that the client emit a behavior that more closely resembles the terminal response.
These behaviors that more and more closely resemble the terminal response are considered
intermediate steps. The fifth step in using shaping is to identify responses that might
constitute intermediate steps between the current response and the final performance. In other
words, what behaviors will be established along the path to the final performance? As each
one is mastered, the criterion for reinforcement is increased to include the next intermediate
behavior. This gradual process continues until the final performance is achieved. It is often
helpful for the practitioner to outline a possible list of component steps in advance. Such a list
is useful in making sure that the practitioner is fully familiar with the dimensions of response
class of interest and provides the practitioner with a crude assessment of how far the client has
to progress to achieve the terminal behavior and how long the intervention might take. Once
the list of potential intermediate steps is complete, however, it is useful to remind oneself of
the maxim The client is always right. In other words, proceed according to the child's
behavior, not the list.
The strategy can be summarized as follows:
1.
2.
3.
4.
5.

Assess to determine the appropriate role of shaping in the intervention.

Determine the reinforcers.
Identify the terminal response class.
Identify and then implement differential reinforcement for the initial response.
Identify and then implement differential reinforcement progressively across the
intermediate responses until the terminal response is reached.

RESEARCH BASIS
The available research suggests that shaping is often most effectively used in combination
with manipulations of antecedent conditions. There are several reasons why this is so. First, a
response cannot be reinforced until it occurs. Thus, shaping requires identification of a
response that the child is currently engaging in that can be used as the starting point.
However, rather than simply waiting for this initial response (e.g., grasping a crayon in
teaching coloring) to spontaneously occur, the practitioner may need to alter the environment
(e.g., by placing the child in a high chair with many crayons before the child, but no other
toys), model the desired response (e.g., by demonstrating to the child picking up the crayon)
or provide a physical or vocal prompt (e.g., by guiding the child's hand to the crayon or by
asking the child to pick up the crayon) to make the response more likely. Once the response
occurs, reinforcement can then be delivered. Prompting and modeling are provided as
necessary at the initiation of each new step but are then gradually withdrawn (i.e., faded) such
that before progressing to the next step, the response is performed without the prompt. This
combined approach appears to expedite the learning process and ensures that the child
receives a relatively high rate of reinforcement while making relatively few errors.
The second reason shaping is often combined with manipulation of antecedents is that for
many responses, not only the occurrence, but also the timing and/or placement of the
response, is important. In these instances, shaping is combined with stimulus discrimination

procedures. Here the developing response class that is targeted via the shaping procedure is
only reinforced in the presence of a discriminative stimulus (and not in its absence). For
instance, in teaching a child to color, only marks of sufficient force (the behavior that is being
shaped) that occur on paper (the discriminative stimulus) are reinforced, while attempts to
mark on walls, tables, clothing, and so on are not reinforced.
Finally, putting it all together, discriminative stimuli, prompts, and shaping are often
simultaneously used. In teaching a child to label (i.e., tact) an apple, the child is shown an
apple (the discriminative stimulus), the practitioner provides a vocal prompt apple, and the
child says appa, which is reinforced. Before concluding, the child's responding will need to
more closely approximate standard usage of apple (the terminal response) and occur
without the formal prompt and only in the presence of apples (and not, say, bananas).

RELEVANT TARGET POPULATIONS, TARGET

BEHAVIORS, AND EXCEPTIONS
Shaping is most well recognized as a component of applied behavior analysis interventions,
which have been widely used in the treatment of children with mental retardation, brain
injury, autism, or other developmental disabilities. The range of target behaviors for which
shaping (in combination with prompts and/or discriminative stimuli) has been used is quite
large, and includes prerequisite skills for learning (e.g., remaining seated, making eye contact,
speaking with sufficient voice volume), language skills(e.g., speech acquisition), social skills
(e.g., sharing, attending to others and getting others' attention, shaking hands), motor skills
(e.g., coordinated hand movements), self-care skills (e.g., self-feeding, toileting, contact lens
use), life skills (telephone use, bed making), and other adaptive behaviors (e.g., school
attendance). This list further illustrates how shaping can be employed across response
dimensions, including different topographies, forces/intensity levels, and amounts/durations.
Shaping is also often used as a component of treatment with more traditional outpatient
populations (e.g., those often diagnosed with learning disorders, attentiondeficit/hyperactivity disorder [ADHD], depression, anxiety, or oppositional behavior). These
children generally have established functional verbal repertoires and are often somewhat
older. As such, aspects of shaping (i.e., treatment goals targeting progressive improvements)
are often used in conjunction with verbal instructions specifying relevant natural and/or any
added contingencies, what is to be done, and when it should occur. For instance, as part of the
treatment for a child diagnosed with separation anxiety disorder, the child might be told that
he or she will receive stickers on a chart (that when completed can be exchanged for a larger
reward) for greater and greater periods of time playing with a babysitter, or alternative
caretaker, in the absence of the primary caretaker. Similarly, when therapeutic homework
practice is sequenced well, it encourages the client to engage in a challenging behavior but
does not request so much as to lead to failure. Shaping can also occur directly in the clinical
setting, such as when the therapist attempts to reinforce small improvements as they are
observed in session. For instance, reinforcing increasingly assertive behaviors emitted by a
generally inhibited adolescent.
An enormous variety of response classes can be, and are, developed and refined via formal
and informal shaping. However, there are some limitations to what shaping can accomplish.
First, shaping applies to operant, not respondent, behavior. As such, while a snake-phobic
child might be gradually taught snakehandling skills and to hold a snake for progressively

longer intervals, accompanying decreases in emotional responses (e.g., anxiety) are not the
result of shaping, but respondent extinction. Second, for shaping to be effective, the child
must have the prerequisite physical capabilities for engaging in the terminal response. There
also appears to be a pragmatic upper limit as to what typical shaping procedures can produce.
For instance, we would be unlikely to be able to routinely shape children to develop the
basketball skill of Michael Jordan. This pragmatic upper limit appears determined by a
number of factors, including (but not limited to) child physical characteristics, the current
state of knowledge in shaping, and the amount and duration of shaping for one particular
response class compared to others.

COMPLICATIONS
A primary complication associated with shaping involves the failure to make progress toward
the terminal behavior. In determining the rate of progress, it must be recognized that even
though the steps may seem logical, straightforward, and nicely sequenced to the practitioner,
that does not mean that the child will progress as such. Rather, progress may at times appear
erratic and better characterized as periods of fast forward and then rewind. In the former, it is
important not to discontinue shaping too early when the terminal behavior may be especially
fragile. In the latter, it may be important to temporarily move back to an established step. In
addition, there are times when progress plateaus. When there is a sustained lack of progress,
the practitioner may be advancing too rapidly or in too large increments. As such, breaking
the intermediate levels down into smaller steps or adding prompts can help remedy the
problem.
The need for creative adjustments during the procedure might also reflect the difference
between the contingency described by the practitioner and the functional contingency
experienced by the child. That is, even though the practitioner may be very clear about what
behavior produced the reinforcer, the contingency experienced by the child may be different.
This is most likely when there is a delay between the response and the reinforcer such that the
child may emit some other behavior prior to reinforcer delivery. In addition, there may be a
loss of reinforcer effectiveness. For instance, if food reinforcement is being used, the child
may become satiated over the course of a long training session. Given concerns about
reinforcer timing and effectiveness, when possible, it is useful to arrange the procedure such
that control can be by, or transferred to, natural reinforcers or conventional social reinforcers,
which do, or can, immediately follow the response and may be less likely to lose their
effectiveness.
There are two additional potential complications that warrant mention. First, there are times
when it is difficult to implement shaping because other incompatible response classes may be
automatically reinforcing or unable to be put on extinction. For instance, it would be difficult
to reinforce coloring if breaking the crayons or throwing the crayons is very reinforcing.
These alternative behaviors are incompatible with coloring, and the reinforcement they
produce is readily available and difficult to put on extinction, because in order to color, the
child must have access to the crayons. Modeling, verbal, and physical prompts, along with
other contingency management strategies, may be required in these situations. A second
complication involves the maintenance and generalization of the shaped behavior. Training in
the natural setting where the behavior is to occur, using natural or conventional social
reinforcers that will be maintained after training, and establishing fluent performance before
discontinuing shaping, can facilitate maintenance and generalization.

CASE ILLUSTRATION
Deb was a 13-year-old female diagnosed as moderately mentally retarded. She lived with
her parents and attended a special education classroom at the public middle school. Deb's
parents always had a difficult time getting Deb to complete a morning hygiene routine (e.g.,
taking a shower, brushing teeth, putting on deodorant, applying lotion to arms and legs, and
combing/brushing hair). Now that she was entering adolescence, it was more important
socially, and for encouraging her development of independence, that she routinely engage in
this class of hygienic behaviors. Assessment materials suggested that Deb has the relevant
physical skills to complete the routine, as she could complete each specific response on
separate occasions (e.g., she might brush her teeth but do none of the others). In addition, she
appeared able to sequence the responses that are part of a behavioral chain (e.g., the
component responses in taking a shower or brushing her teeth). However, simply posting a list
of what she was to accomplish each morning and attempting to reinforce completion of the
entire routine have not been effective interventions. Thus, Deb's noncompliance appeared, at
least in part, a result of a response amount deficittoo much was being asked of her.
Based on a preference assessment in which Deb was presented with a series of two items and
asked to choose one, glitter pens were identified as especially reinforcing for her. Thus, a
reinforcement procedure was implemented in which Deb earned stickers toward the purchase
of new glitter pens. In addition, in an attempt to make the hygiene behaviors as naturally
reinforcing as possible, Deb was taken to a local retail store with her parents and allowed to
select the toothpaste and brush, deodorant, shampoo, lotion, and hair brush and comb of her
choice.
Deb's parents defined the terminal response class as completion of an established hygiene
repertoire consisting of the following: taking a shower, brushing teeth, putting on deodorant,
applying lotion to arms and legs, and combing/brushing hair. Baseline assessments revealed
that Deb typically managed to do at least one, occasionally two, of these behaviors, but never
completed all of them independently. Thus, an intervention was developed in which each
morning Deb would be prompted to look at the list of responses and to select one or more to
complete in order to earn a sticker. Completion of any one of the responses was considered as
meeting the initial response requirement. The intermediate responses were defined as
completing two to four items.
After completing the initial response requirement on three consecutive days, the criterion for
reinforcement was increased and Deb was prompted each morning to independently complete
two or more of the responses to earn a sticker. From that day on, Deb earned the sticker only
after completing two responses (completing one behavior was no longer reinforced). After
three successful days, the response requirement was again increased, and Deb was prompted
each morning to independently complete three or more of the responses to earn a sticker. This
process continued until Deb consistently completed all five items of the morning routine. At
this point, prompting Deb to look at the list was faded to increase her independent
engagement in the routine. In addition, as she progressed through the procedure, her parents,
extended family, school staff, and therapists praised relevant aspects of her appearance to
attempt to facilitate transfer to the natural outcome of good hygiene and conventional social
reinforcers. It is hoped that over time these reinforcers will come to maintain the routine, and
use of the stickers can be gradually withdrawn.
Scott T. Gaynor and Jean Clore

Further Reading

Entry Citation:
Gaynor, Scott T., and Jean Clore. "Shaping." Encyclopedia of Behavior Modification and
Cognitive Behavior Therapy. 2007. SAGE Publications. 15 Apr. 2008. <[Link]