Transcript Joanna LM schedules of reinforcement

Introduction to Schedules of
Reinforcement
Who am I?
• Joanna Lomas Mevers, PhD, BCBA
• 13 years of experience conducting
research and clinical treatment using the
principles of applied behavior analysis to
treat a variety of problem behaviors
observed in children
Applied Behavior Analysis
• Baer, Wolf, & Risley, (1968)
– Applied
– Behavioral
– Analytic
– Technological
– Conceptually Systematic
– Effective
– Generalizable
Applied Behavior Analytic
Research
– Applied
• “Teaching teenagers with autism to seek
assistance when lost”
• “Teaching safety skills to children to prevent gun
play”
• “Delay discounting by pathological gamblers”
• “Social antecedents of children’s eyewitness
testimony”
• “A half century of scalloping in the work habits of
the United States Congress”
A Day in the Life of a Behavior
Analyst
• Universities
– Professors, researchers
• Education
– Consultants, school psychologists, special educators
• Industry
– Safety/efficiency consultants, sports psychologists, military
consultants
• Private practice psychologists
– Counselors, parent training specialists
• Healthcare
– Pediatric and developmental psychologists
• Miscellaneous
– animal trainers, social behavior specialists
Behavior Problems with a History of Successful Treatment by
Behavior Analytic Interventions
•
•
•
•
•
•
•
•
•
•
•
•
•
Phobias
Obsessive-compulsive disorder
Speech delays
Paraphillias
Self injury
Aggression
Property destruction
Pica
Conduct disorders (i.e., juvenile delinquency)
Eneurisis/econpresis
Inattention/impulsivity
Feeding disorders
Rumination
A few of the tools?
•
•
•
•
Schedules of Reinforcement
Punishment
Motivation
Stimulus Control
Why Study Schedules of
Reinforcement??
• As behavior analyst this is one of the most
important tools we have to change
behavior
Patterns of Behavior
• “The outstanding characteristic of
operant behavior is that it can be
differentiated in form and in temporal
patterning by consequent events.”
-Morse, 1966
Patterns of Behavior
• Behavior is very, very complex---but:
– Behavior is not chaotic or without causes
– This means there can be a science of
behavior
– There are orderly patterns to behavior
– These patterns are induced, in part, by
contingencies
CUMULATIVE STUDY TIME FOR TEST
T
Operant Conditioning
W
Th
F
S
Su
M
T
Important Variables to Consider
• Motivation
• Individual differences
• Reinforcement
• Punishment
• Stimulus Control
• Discrimination
• Generalization
• Match–to-sample
Three-Term Contingency
1. Operant class (Ro)
2. Discriminative Stimulus (SD)
3. Reinforcing Consequences (SR)
Schedule Classification
• Response dependent and independent
• Which one is Pavlovian?
Response Independent
• Fixed-time
• Random time
• Extinction
Response Dependent Schedules
• Ratio
• Interval
• DRO
Fixed Ratio (FR) Schedule
• Definition: A response is reinforced when
a fixed number of responses have been
emitted since the last reinforcer.
Fixed-ratio cumulative record:
FR PERFORMANCE
100
\
\
80
CUMULATIVE RESPONSES
\
\
\
\
\
\
\
\
\
\
40
\
20
\
100
200
TERMINAL RUN
REINFORCER
\
\
\
\
\
\
\
\
0
\
\
\
\
\
60
\
300
PAUSE AFTER RFT
\
\
\
400
TIME
Simple schedules and feedback
functions
500
600
18
Fixed Ratio (FR) Schedules:
Independent Variables
• The terminal run -- insensitive
• The pause-after-reinforcement -- sensitive
to many independent variables –
– Deprivation
– Punishment
– Ratio requirements
Fixed Ratio (FR) Schedules
• Extinction after FR training:
– Burst-and-break pattern, pauses getting
longer, runs getting shorter
Variable Ratio (VR) Schedule
• Definition: A response is reinforced when
a number of responses that varies around
some average have been emitted since
the last reinforcer
• Range of responses varies:
– VR 30 could be M = 30 (1-60) or M = 30 (28-32)
Variable-ratio cumulative record:
VR PERFORMANCE
100
\
CUMULATIVE RESPONSES
\
80
\
REINFORCERS
60
\\
40
\
20
\
0
100
200
300
TIME
400
500
600
22
Variable Ratio (VR) Schedules:
Independent Variables & Extinction
• Insensitive to most independent variables
– Performance breaks up under
• Very low levels of deprivation
• Very high response requirements
• Extinction after VR training:
– Burst-and-break pattern
– VR performance takes longer to extinguish
than FR performance
Not a VR schedule
• Some say playing slot machines and other
gambling activities are reinforced on a VR
schedule
•This is “the gambler’s
fallacy – each
response increases
the probability of the
next response being
reinforced
•These activities
actually operate on a
random ratio schedule
Fixed Interval (FI) Schedules
• Definition: A response is reinforced when
a fixed period of time has elapsed since
the last reinforcer (or event).
Scalloped pattern – low rate after reinforcers,
accelerates to a high rate before the next reinforcer.
Fixed-Interval (FI)1-min.
Fixed Interval (FI) Schedules:
Independent Variables & Extinction
Independent variables
• Quite sensitive to many independent variables -drugs, deprivation, punishment, etc
Extinction
• Steady decrease in response rate. Overall,
extinction takes much the same time as an FR
performance with the same reinforcer rate.
Variable Interval (VI) Schedules
Definition: A response is reinforced when a
period of time that varies around some
mean has elapsed since the last reinforcer
(or event).
High, constant rate of responding with few pauses. But
not as high as VR
VI-SCHEDULE PERFORMANCE
100
\
\
\
CUMULATIVE RESPONSES
80
\
\
60
\
\
\
40
\
20
0
0
100
200
300
TIME
400
500
600
Variable Interval (VI) Schedules:
Independent Variables & Extinction
Independent variables
• Quite sensitive to various independent variables
-- reinforcer rate, reinforcer magnitude,
deprivation, punishment, and drugs.
Extinction
• Smooth and progressive decline in response
rate until the behaviour is eliminated (just like
extinction after FI training). But VI performance
takes a longer time than FI to extinguish.
Differential Reinforcement of
Other Behavior (DRO)
• Definition: A reinforcer is delivered when a defined
response has not been emitted for a fixed period of time.
• Note that the delivery of the reinforcer is not contingent
on a particular response being emitted.
Differential Reinforcement of
Other Behavior (DRO)
Example: A child is given a sweet after not
showing any tantrums for a minute (DRO 60 s). If
no more tantrums are emitted, a sweet is given
every minute.
• Notice that, in DRO schedules, the contingency
is on the complement of the defined response -on not lever pressing, or not key pecking, or not
emitting tantrums.
Differential Reinforcement of
Other Behavior (DRO)
• 2 types:
– Resetting: each time the behavior that is targeted for reduction
occurs the timer resets
– Non-resetting: the timer runs its course each time, and the
behavior is either reinforced (or not) depending on whether it has
occurred in that interval
Differential Reinforcement of
Other Behavior (DRO)
• Comment: DRO schedules are sometimes also called
omission training and (really confusingly) negative
punishment schedules. Avoid the latter two.
• Used extensively to eliminate unwanted behavior:
• quicker than simply extinguishing the behaviour
• extinction sometimes increases the unwanted
behavior briefly, and can also lead to aggression
• Ethically more acceptable than punishment
procedures (although they may take longer to have
the desired effect).
Differential Reinforcement of
Other Behavior (DRO)
• Comment #2: DRO schedules are not actually
reinforcing the absence of a behavior (you can’t reinforce
nothing).
• Rather, any other behavior that happens to be occurring
when the contingency is met increases, displacing the
behavior that is targeted for decrease
• In this way, what type of schedule is most similar to a
DRO?
Which Would You Use?
• Out of FR, VR, FI, VI
• To get the highest rate of responding?
• When you want a behavior to be resistant
to extinction?
• For the fastest extinction?
• Teach a new response?