Differential
Reinforcement
is a behavioral procedure indented to
increase the frequency of a desired
behavior and to decrease the frequency of
undesirable behaviors. The result is an
increase in the future probability of the
desirable behavior.
All applications of differential reinforcement
entail reinforcing one response class and
withholding reinforcement for another response
class.
When used as a reductive procedure for
problem behavior, differential reinforcement
consists of two components: (a) providing
reinforcement contingent on either the
occurrence of a behavior other than the
problem behavior or the problem behavior
occurring at a reduced rate, and (b) withholding
reinforcement as much as possible for the
problem behavior.
Types of Differential Reinforcement
DRA (differential reinforcement of
alternative behavior)
DRO (differential reinforcement of other
behavior)
DRL (differential reinforcement of low
rates of responding)
DRA - Differential Reinforcement of
Alternative Behavior
Reinforcer is delivered for desirable
behavior
Extinction for the problem behavior
It has the dual effects of weakening the
problem behavior while simultaneously
strengthening acceptable behaviors that
are either incompatible with or an
alternative to the targeted problem
behaviors.
Examples of DRA
A child’s good table manners are reinforced with
praise and by passing the requested food. The
child’s bad table manners are extinguished by
ignoring and not passing food.
Punching the correct code numbers in the ATM
is reinforced with cash. Punching the wrong
numbers in the ATM is extinguished by getting
no cash.
Reinforce playing or sharing to replace fighting
in children.
Reinforce lying quietly in the dentists chair to
replace disruptive behavior.
Reinforce an assertive response to replace
aggressive behavior.
reinforced the in-seat behavior of a
preschool student with hyperactivity, out-
of-seat behavior decreased markedly.
When to Use DRA
1. You want to increase a desirable
behavior and/or decrease undesirable
behaviors.
2. The desirable behavior already occurs at
least occasionally.
3. You have a reinforcer you can use.
How to choose reinforcers
Ask
Observe
Test (a preference assessment procedure)
To determine that the item did in fact
function as a reinforcer, you would deliver it
contingent on a behavior and show that the
behaviour increased. This process is a
reinforcer assessment.
In the single stimulus assessment
procedure, each potential reinforcer is
presented (i.e., put on a table in front of
the child), one at a time, to see whether
the individual approaches the stimulus or
not.
In the paired stimulus assessment
procedure (also called a forced choice or
paired choice procedure), two potential
reinforcers are presented to the individual
and the researcher records which stimulus
the individual approaches.
In the multiple stimulus assessment
procedure, an array of potential reinforcers is
presented to the individual (for example, 8
stimuli are presented on a table
in front of the person), and the researcher
records which potential reinforcer the
individual approaches or chooses first. This
stimulus is then removed from the
array and the researcher records which
stimulus the individual chooses next.
Using DRA
1. Define the target behavior to increase and the
undesirable competing behavior(s) to decrease
2. Identify the reinforcer for the problem behavior
3. Choose a reinforcer for the desirable behavior
4. Reinforce desirable behavior immediately and
consistently
- prompt the desirable behavior if necessary
- the desirable behavior should require less
response effort than the problem behavior
5. Eliminate (or minimize) the reinforcer for the
undesirable behavior(s)
6. Use intermittent reinforcement for maintenance
Considerations in the Use of DRA
Consider establishing operations to make
the reinforcer more potent
Consider the use of rules or instructions
Fade artificial reinforcers to natural
reinforcers to help maintain the behavior
Consider using the Premack principle
Differential Negative Reinforcement
of Alternative Behavior
Desirable behavior produces escape or
avoidance of aversive stimulus
Undesirable behavior does not produce
escape/avoidance
Example
When the teacher asked Jason to do his
schoolwork (e.g., complete problems in his
workbook), he often slammed his fists on
his desk and rocked back and forth
violently in his seat. At this, the teacher
usually let Jason take a break and sit in a
chair by himself at the back of the room
until he calmed down. Because this
behavior occurred four or five times every
day, Jason was not getting much
schoolwork done.
Because escape from his schoolwork was
reinforcing his outbursts, she did not let
him escape: He could not get out of his
seat and sit in the back of the room when
he had an outburst. Instead, he had to
stay in his seat and, when he calmed
down, he still had to do the workbook
problem. In this way, doing his workbook
problems resulted in reinforcement and
having outbursts did not result in
reinforcement.
Variations of DRA
DRI = Differential reinforcement of
incompatible behavior
DRC = Differential reinforcement of
communication (functional communication
training)
DRI = Differential reinforcement
of incompatible behavior
The alternative behaviour is physically
incompatible with the problem behavior
and, therefore, the two behaviours cannot
occur at the same time.
Example
For example, if the problem behavior is
head-slapping, in which individuals slap
themselves on the side of the head with their
hands, any alternative behavior involving the
use of the hands would be an incompatible
behavior. Playing with toys or completing
tasks that involve the manipulation of
materials with their hands would be
examples of incompatible behaviours that
could be reinforced to replace the head-
slapping in a DRI procedure.
DRC = Differential reinforcement of
communication (functional communication
training)
n this procedure, the individual with the
problem behavior learns to make a
communication response that is
functionally equivalent to the problem
behavior. When the communication
produces the same reinforcing outcome as
the problem behavior, there is no longer
any reasonfor the problem behavior to
occur.
Example of DRC
Participants to escape from a task for 30
seconds contingent on using an alternative
response (e.g., giving the therapist a card
with “BREAK” printed on it, or saying “no”)
as an alternative to their aberrant
behavior. The researchers also provided
praise for using the alternative responses.
DRO - Differential Reinforcement
of Other Behavior
Also called differential reinforcement of zero rate
of behavior
Reinforcer is delivered for the absence of the
problem behavior in intervals of time
Intervals are chosen based on the baseline level
of the problem behavior
Extinction for the occurrence of the problem
behavior
If the problem behavior occurs, the interval is
reset
Examples of DRO
Attention delivered every 15 sec without
SIB
Breaks from academic tasks delivered
every 20 minutes for the absence of
problem behaviors
Access to a favorite toy given each ½ hour
for the absence of fighting with siblings
as a group contingency to decrease the
disruptive classroom behaviors of third-grade
students. A kitchen timer was set to 5-minute
intervals and continued to run as long as no
disruptive behaviors occurred. If any student
engaged in disruptive behavior at any time
during the 5- minute interval, the timer was
reset and a new 5-minute interval began.
When the timer signaled the end of a 5-
minute interval without disruptive behavior,
the class was awarded 1 minute of free time,
which they accumulated and used at the end
of the class period.
Whole Interval vs Momentary DRO
Whole interval DRO - the problem must be
absent for the whole interval for
reinforcement (referred to simply as DRO)
Momentary DRO - the problem must be
absent at the end of the interval for
reinforcement
Implementing DRO
1. Identify the reinforcer for the problem behavior
2. Identify the reinforcer to use in DRO
3. Choose the DRO interval
4. Use extinction for the problem behavior
(or minimize the reinforcement for the problem)
5. Deliver the reinforcer for the absence of the
problem behavior in each interval
6. The occurrence of the problem behavior resets
the interval for delivery of the reinforcer
7. Gradually increase the DRO interval
8. Use instructions when applicable
DRL - Differential Reinforcement of
Low Rates of Responding
Used to get a behavior to occur less (not
necessarily to eliminate the behavior)
Two types of DRL
- Full session DRL
- Spaced responding DRL
Full Session DRL
The reinforcer is delivered when fewer
than a specified number of responses
occurs per time period (session)
The reinforcer is delivered at the end of
the session
The timing of responses in the session is
not important
Spaced Responding DRL
The reinforcer is delivered for a response
when it is separated from the previous
response by a specified interval of time
If a response occurs before the end of the
interval, the interval is reset
Individual responses are reinforced when
the IRT is greater than X
The timing of responses is important
Examples of DRL
Full session DRL
Reward given for smoking fewer than 5
cigarettes per day
Dessert given if the child gets up from the table
fewer than 3 times during supper
Spaced responding DRL
Child called on to answer in class only if it has
been 10 minutes since he last raised his hand
Person with ID allowed to take a bite of food only
if 15 sec since last bite of food
Implementing DRL
1. Is DRL the appropriate procedure?
- Do you want to decrease but not eliminate
the behavior?
2. Determine the acceptable level of the behavior
3. Decide on full session DRL or spaced
responding DRL
4. Inform the client of the criterion for
reinforcement
5. Use a procedure to give the client feedback on
number of responses or timing of responses
6. Use intermediate goals if necessary
Comparing DRO and Spaced
Responding DRL
DRO - at the end of the interval, the
reinforcer is delivered for the absence of
the behavior
DRL - at the end of the interval, the
reinforcer is delivered for the first response
In DRO and DRL, a response before the
end of the interval resets the interval
